JP2013152692A - Database construction device, database construction method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that two or more relevant technical terms cannot be extracted and information other than technical terms cannot be extracted in a conventional technical term extraction device.SOLUTION: A database construction device comprises: a table storage part for storing a first table including one or more attribute value groups each having m (m≥3) attribute values different in attribute, and two or more second tables including one or more attribute value groups each having n (n≤m-1) attribute values different in attribute; a rule storage part for storing a rule for acquiring one or more attribute value groups each having n attribute values from a Web page; an acquisition part for acquiring one or more attribute value groups each having n attribute values from a Web page according to the rule to add the one or more attribute value groups to the second table; and a coupling part for coupling two or more attribute value groups which two or more second tables have, generating an attribute value group having m attribute values, and adding it to the first table. All of information on the Web can be made into a database by the above-described database construction device.

Description

  The present invention relates to a database construction device for constructing a database.

  2. Description of the Related Art Conventionally, a technical term extraction device that extracts technical terms from Web pages has been developed (see Patent Document 1).

JP 2008-257511 A

  There is a huge amount of all kinds of information on the Web. However, the information is not structured and the necessary information usually exists in a dispersed state. A conventional technical term extraction device cannot construct a database by structuring distributed information.

  The database construction apparatus according to the first aspect of the present invention includes a first table having one or more attribute value sets having m attribute values (m ≧ 3) having different attributes, and n items having different attributes (n ≦ m− 1) One or more attribute values having n attribute values from one or more Web pages, and a table storage unit storing two or more second tables having one or more attribute value sets having the attribute value 1) One or more rules for acquiring a set are stored in association with the second table, and a one or more attribute value sets having n attribute values from one or more Web pages according to the rules Is obtained, and the acquisition unit for appending to the second table and the two or more attribute value sets of the two or more second tables are combined to generate one or more attribute value sets having m attribute values. , A database construction device comprising a coupling unit to be added to the first table.

  With such a configuration, a structured database can be constructed from one or more Web pages.

  The database construction device according to the second aspect of the invention provides the rule storage unit with one or more attribute value sets having m attribute values from one or more Web pages. Are stored in association with the first table, and the acquisition unit acquires one or more attribute value sets having m attribute values from one or more Web pages according to the rule, It is a database construction device that adds to the table.

  With such a configuration, the attribute value set included in the first table can be acquired.

  Further, in the database construction device of the third invention, the second table having a unique identifier indicating whether or not the attribute is unique is stored in the table storage unit with respect to the first invention. Part is an attribute value set of each of two or more different second tables, has two or more attributes that have the same attribute value of the same attribute, and that the attribute is indicated by a unique identifier This is a database construction apparatus that combines a value set with an attribute value as a key, generates one or more attribute value sets having m attribute values, and adds them to the first table.

  With such a configuration, two or more attribute value sets can be combined to generate an attribute value set having three or more attributes.

  Further, in the database construction device according to the fourth aspect of the present invention, with respect to the first aspect, the combining unit is an attribute value set possessed by each of three or more different second tables, and the same attribute having the same attribute This is a database construction device that combines an attribute value set having two or more values with the attribute value as a key, generates one or more attribute value sets having m attribute values, and adds them to the first table.

  With such a configuration, an attribute value set having three or more attributes can be generated from three or more attribute value sets having two or more attributes.

  In the database construction device according to the fifth aspect of the present invention, the Web page has a Web page identifier for identifying the Web page, and the acquisition unit has one or more attributes. When acquiring a value set, a database construction apparatus that also acquires a Web page identifier of a Web page and adds the Web page identifier to the first table or the second table in association with one or more attribute value sets is there.

  With such a configuration, the acquisition source of the attribute value set can be managed.

  In the database construction device according to the sixth aspect of the present invention, when the acquisition unit acquires one or more attribute value sets, the acquisition unit also acquires the score of the attribute value set for any of the first to fifth aspects of the invention. In addition, the score is added to the first table or the second table in association with one or more attribute value sets, and the combining unit has a second table with a higher score corresponding to a predetermined condition. This is a database construction device that combines two or more attribute value sets to generate one or more attribute value sets having m attribute values and appends them to the first table.

  With such a configuration, an attribute value set with a high score can be generated.

  In addition, the database construction device of the seventh invention further includes a score storage unit that stores one or more scores associated with the Web page identifier, as compared with the fifth or sixth invention, and an acquisition unit When acquiring one or more attribute value sets, it also acquires a Web page identifier that the Web page has, acquires a score corresponding to the Web page identifier from the score storage unit, and associates it with one or more attribute value sets The database construction device additionally writes the Web page identifier and the score to the first table or the second table.

  With such a configuration, an attribute value set with a high score can be generated.

  Further, the database construction apparatus of the eighth invention further includes a dictionary storage unit that stores one or more dictionaries that are a set of one or more attribute values for each attribute, as compared with the sixth invention. When acquiring one or more attribute value sets, the attribute value is an attribute value of the attribute value set and the attribute value of the same attribute as the attribute value attribute. It is a database construction device that calculates and adds a score to the first table or the second table in association with one or more attribute value sets.

  With such a configuration, an attribute value set with a high score can be generated.

  The database construction apparatus according to the ninth aspect of the invention is a character string in which one or more attribute value sets stored in a table storage unit are present from one or more Web pages with respect to the first to eighth aspects of the invention. Is a database construction device that further includes a rule acquisition unit that acquires a character string other than the attribute value set among the character strings as a rule and accumulates it in the rule storage unit.

  With such a configuration, the number of rules can be increased.

  In the database construction device according to the tenth aspect of the present invention, in the ninth aspect, the rule acquisition unit has a set of attribute values used when acquiring the rule for the acquired one or more rules. It is a database construction device that accumulates in a rule storage unit in association with a table or a second table.

  With such a configuration, rules can be managed for each attribute corresponding to an attribute value that can be acquired.

  In the database construction device according to the eleventh aspect of the invention, in contrast to the first to tenth aspects, the joining unit is an attribute value set having two or more second tables stored in the table storage unit. Yes, two or more attribute value sets of the attribute value set acquired by the acquisition unit or the attribute value set generated by the attribute value generation unit are combined to generate one or more attribute value sets having m attribute values. This is a database construction device that is added to the first table.

  With such a configuration, an attribute value set having a larger number of attributes than the number of attributes can be generated using an attribute value set having a smaller number of attributes.

  According to the database construction device or the like according to the present invention, a structured database can be constructed from one or more Web pages.

Block diagram of database construction apparatus in embodiment 1 Block diagram of the database construction device in the same embodiment Block diagram of trademark infringement detection apparatus in the embodiment The block diagram of the trademark similarity judgment part of the trademark infringement detection device in the embodiment The flowchart which shows an example of operation | movement of the company acquisition part of the database construction apparatus in the embodiment The figure which shows an example of the data stored in the company rule database in the embodiment The flowchart which shows an example of operation | movement of the goods acquisition part of the database construction apparatus in embodiment of this invention in the embodiment The figure which shows an example of the data stored in the goods rule database in the embodiment The flowchart which shows an example of operation | movement of the unique name acquisition part of the database construction apparatus in the embodiment The figure which shows an example of the data stored in the proper name rule database in the embodiment The flowchart which shows an example of operation | movement of the knowledge acquisition part of the database construction apparatus in the embodiment The figure which shows an example of the data stored in the knowledge proliferation rule database in the embodiment The flowchart 1 which shows an example of operation | movement of the rule learning part of the database construction apparatus in the embodiment The flowchart 2 which shows an example of operation | movement of the rule learning part of the database construction apparatus in the embodiment The flowchart which shows an example of operation | movement of the database construction apparatus in the embodiment The flowchart which shows an example of operation | movement of the trademark infringement detection apparatus in the embodiment The figure which shows the example of a display of the trademark infringement detection apparatus in the form of the embodiment The figure which shows an example of the external appearance of the computer system in the embodiment The figure which shows an example of a structure of the computer system in the embodiment Block diagram of trademark infringement detection apparatus in Embodiment 2 Block diagram of database construction device 5 in Embodiment 3 The flowchart explaining the whole operation | movement of the database construction apparatus 5 in the embodiment Flowchart for explaining processing for acquiring the same attribute value set Flowchart explaining generation processing of same attribute value set Flowchart explaining the rule acquisition process Diagram showing an example of the same rule The figure which shows the example of the same Web page The figure which shows the example of the same Web page The figure which shows the example of the 1st table and the 2nd table The figure which shows the example of the 1st table and the 2nd table The figure which shows the example of the 1st table and the 2nd table Figure showing an example of the second table Figure showing an example of the second table Figure showing an example of the second table Figure showing an example of the second table Figure showing an example of the second table

  Hereinafter, embodiments of a database construction device and the like will be described with reference to the drawings. In addition, since the component which attached | subjected the same code | symbol in embodiment performs the same operation | movement, description may be abbreviate | omitted again.

(Embodiment 1)
FIG. 1 is a block diagram of a database construction device 1 according to the present embodiment. The database construction device 1 includes a company database 101, a product database 102, a unique name database 103, a company product database 104, a company unique name database 105, a product unique name database 106, a company product unique name database 107, a company rule database 108, a product rule. A database 109, a unique name rule database 110, a knowledge multiplication rule database 111, a company name acquisition unit 112, a product name acquisition unit 113, a unique name acquisition unit 114, a knowledge acquisition unit 115, and a rule learning unit 116 are provided.

  The company database 101 can store one or more company names. The company name is a name having a legal personality such as a stock company, a limited liability company, a public interest corporation, an incorporated association or an independent administrative corporation that provides goods or services, and an individual name.

  Further, the company name may be information for specifying the company name. The company database 101 may store not only the company name but also information on the document from which the company name has been acquired, the number of acquisitions of the company name, the acquisition ratio, and the like. The information for identifying the company of the company name may be any information that can uniquely identify the company. For example, it may be information indicating a home page address, an address, a telephone number, or the like. The information of the document from which the company name has been acquired is, for example, the name of the document, the number of pages, information indicating the URL of the Web site, or the text before and after the company name, or the HTML tag structure.

  Storage is a concept that includes long-term storage using a non-volatile recording medium and temporary storage using a volatile recording medium. Hereinafter, the concept of storage in the present invention is the same.

  Further, the process of storing the company name in the company database 101 does not matter. For example, a company name may be stored in the company database 101 via a recording medium, or a company name transmitted via a communication line or the like may be stored in the company database 101. Alternatively, the company name input via the input device may be stored in the company database 101. The company database 101 is preferably a non-volatile recording medium, but can also be realized by a volatile recording medium.

  A product database 102, a unique name database 103, a company product database 104, a company unique name database 105, a product unique name database 106, a company product unique name database 107, a company rule database 108, a product rule database 109, a unique name rule, which will be described later. Information is stored in each component of the database 110, the knowledge multiplication rule database 111, the product similar group code correspondence database 21, the company similar group code unique name database 22, the database 31, the database 401, and the product code management means 4021. Since the process and the means for realizing them are the same as those of the company database 101, description thereof is omitted.

  The product database 102 can store one or more product names. The product name is the name of a product or service sold or provided by a corporation and an individual. Specifically, “chocolate”, “computer”, “delivery”, “moving service”, and the like. The names of goods and services mentioned here are not necessarily the names of goods and services described in the classification of goods and services (hereinafter referred to as similar group code table) stipulated in Article 6 of the Enforcement Regulations of the Trademark Law. It does not have to comply.

  Further, the product name may be information for specifying the product name. The product database 102 may store not only the product name but also the information of the document from which the product name is acquired, the number of acquisitions of the product name, the acquisition ratio, and the like. The information specifying the product with the product name may be anything as long as it is information that can uniquely specify the product. For example, it may be information indicating a home page address, an address, a telephone number, or the like. The information of the document from which the product name has been acquired is, for example, the name of the document, the number of pages, information indicating the URL of the Web site, or the text before and after the product name, the HTML tag structure, or the like.

  The unique name database 103 can store unique names of one or more products. The unique name of the product is a name that can identify the product and service. Specifically, “Yukimi Daifuku (registered trademark)” “Takkyubin (registered trademark)” “PatentSearch Assistant (registered trademark)” and the like. In addition, the proper name of the product referred to here may or may not be a trademark registered.

  Further, the unique name of the product may be information for specifying the unique name of the product. The unique name database 103 may store not only the unique name of the product but also the information of the document that acquired the unique name of the product, the number of acquisitions of the unique name of the product, the acquisition ratio, and the like. The information that identifies the unique name of the product may be any information that can uniquely identify the unique name of the product. For example, it may be a trademark application number or a publication number. The information of the document from which the product unique name is acquired is, for example, the document name, the number of pages, information indicating the URL of the Web site, or the text before and after the product unique name, the HTML tag structure, or the like.

  The company product database 104 may store one or more pairs having a company name and a product name.

  The company product database 104 stores not only the company name and the product name but also the information of the document that acquired the company name and product name in the same record, or the number of times the company name and product name are acquired in the same record, the acquisition frequency, Occurrence frequency, the average distance based on the number of characters, the number of morphemes, or the number of words listed between the company name and the product name in the document, or the total or average of the number of acquisitions and appearance frequency values of each rule described later The value calculated by using any method may be used. Etc. may be stored. The information of the document in which the company name and product name in the same record are acquired is, for example, the document name and the number of pages, or information indicating the URL of the website, the text before and after the company name and product name, and the HTML tag Such as structure.

  The company unique name database 105 may store one or more sets having a company name and a product unique name.

  The company unique name database 105 stores not only the company name and the unique name of the product, but also the information of the document that acquired the company name and the unique name of the product in the same record, or the company name and the unique name of the product in the same record. Number of acquisitions, acquisition frequency and co-occurrence frequency, the average number of characters, the number of morphemes, the number of words, etc. written between the company name and the unique name of the product in the document, or the acquisition number of each rule described later Or a value calculated by some method such as the sum or average of the appearance frequency values may be stored. The information of the document in which the company name and the unique name of the product in the same record are acquired is, for example, the name of the document, the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  The product unique name database 106 may store one or more sets having product names and product unique names.

  The company unique name database 105 stores not only the product name and the product unique name, but also the information of the document in which the product name and product unique name in the same record are acquired, or the product name and product unique name in the same record. Acquisition count, acquisition frequency, co-occurrence frequency, average distance between the product name and the product unique name in the document, the number of characters, the number of morphemes, the number of words, etc. Or a value calculated by some method such as the sum or average of the appearance frequency values may be stored. The information of the document in which the product name and the unique name of the product in the same record are acquired is, for example, the document name and the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  The company product unique name database 107 may store one or more sets having a company name, a product name, and a product unique name.

  The company product unique name database 107 includes not only the company name, the product name, and the product unique name, but also the information on the document that acquired the company name, product name, and product unique name in the same record, or the same record. Number of acquisitions of company name, product name and unique name of product, acquisition frequency and co-occurrence frequency, number of characters written between each of company name, product name and unique name of product in document, morpheme number, or An average distance based on the number of words or the like, or a value calculated by using some method such as the total or average of the number of acquired rules and the appearance frequency described later may be stored. The information of the document in which the product name and the unique name of the product in the same record are acquired is, for example, the document name and the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  Further, the company product unique name database 107 may be registered from information on trademarks that have already been filed. In that case, the identification information (for example, registration number) or information such as the registration status may be stored.

  The company rule database 108 may store one or more company rules that are rules for obtaining a company name from a document.

  A document is a collection of one or more sentences. The company name, product name, and product unique name do not necessarily have to be written, and are not only Japanese but also English, Chinese, and all other languages. Alternatively, two or more languages may be mixed. For example, there may be a Web page, HTML, XML, Word file, or the like.

  Obtaining a company name from a document means that a part in which the company name is described is determined from the document and the company name is extracted.

  The company rule is a rule for extracting a company name from a document. For example, it may be a sentence pattern in which the company name can be directly extracted, or a sentence pattern that specifies a sentence in which the company name is described. When the document is a Web page, it may be an HTML tag structure pattern that directly extracts a company name, or an HTML tag structure pattern that specifies a sentence in which the company name is described. . Moreover, the same pattern as the above which cannot extract a company name may be sufficient. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, in the case of a company rule for acquiring a company name described in the “% company name%” portion, “<td> company name </ td> <td>% company name% </ td>”. And so on.

  The company rule database 108 may store not only the company rules but also the number of acquisitions and the acquisition ratio of the company rules, or the targets extracted by the company rules. The target to be extracted by the company rules is when there is a possibility of acquiring multiple types of information, such as when acquiring a company name or when acquiring a sentence including the company name as a result of using the company rule from a document , A value that clearly specifies what to get. Specifically, values such as “company name” or “text” may be stored.

  The merchandise rule database 109 can store one or more merchandise rules that are rules for obtaining a merchandise name from a document.

  Obtaining the product name from the document means that the part in which the product name is described is determined from the document and the product name is extracted.

  The product rule is a rule for acquiring a product name from a document. For example, it may be a sentence pattern from which a product name can be directly extracted, or a sentence pattern that specifies a sentence in which a product name is described. If the document is a Web page, it may be an HTML tag structure pattern that directly extracts a product name, or an HTML tag structure pattern that specifies a sentence in which the product name is described. . Moreover, the pattern similar to the above which cannot extract a brand name may be sufficient. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, in the case of a product rule for acquiring a product name described in the “% product name%” part, “<title> product name list </ title>. *? <Div. *?> <H [ 1-5]. *?> List <h [1-5]> </ div> <ul. *?> (<Li. *?> <A. *?>% Product name% </a> <// </ Li>) + </ ul> ".

  The product rule database 109 may store not only the product rules but also the number of acquisitions and the acquisition ratio of the product rules, or the objects extracted by the product rules. The target of product rule extraction is when there is a possibility of acquiring multiple types of information, such as when acquiring a product name as a result of using a product rule from a document, or when acquiring a sentence containing the product name , A value that clearly specifies what to get. Specifically, a value such as “product name” or “text” may be stored.

  The unique name rule database 110 can store one or more unique name rules, which are rules for acquiring a unique name of a product from a document.

  Obtaining the unique name of the product from the document means that the part in which the unique name of the product is described is determined from the document and the product name is extracted.

  The unique name rule is a rule for acquiring a unique name of a product from a document. For example, the pattern of the text which can extract the specific name of goods directly may be sufficient, and the text pattern which specifies the text in which the unique name of goods is described may be sufficient. If the document is a Web page, it may be an HTML tag structure pattern that directly extracts the unique name of the product, or an HTML tag structure pattern that specifies a sentence in which the unique name of the product is described. There may be. Moreover, the same pattern as the above which cannot extract the proper name of goods may be sufficient. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, in the case of the unique name rule for acquiring the unique name of the product described in the “% unique name%” part, the “<title> product list </ title>. *? <Span. *?> List </ Span> <ul. *?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul>

  The unique name rule database 110 may store not only the unique name rule but also the number of acquisitions and the acquisition ratio of the unique name rule, or the target extracted by the unique name rule. The target to be extracted by the unique name rule is the acquisition of multiple types of information, such as when acquiring a unique name of a product or when acquiring a sentence containing the unique name of a product as a result of using the unique name rule from a document. If there is a possibility, it is a value that specifies what to get. Specifically, a value such as “product unique name” or “text” may be stored.

  The knowledge proliferation rule database 111 can store one or more knowledge proliferation rules that are rules for acquiring information with a larger number of components from a document using information with a smaller number of components.

The acquisition of information with a larger number of components using information with a smaller number of components is the following information.
-Using the information stored in the company database 101, obtaining information stored in the company product database 104, the company unique name database 105, and the company product unique name database 107-Stored in the product database 102 The information stored in the company product database 104, the product unique name database 106, and the company product unique name database 107 is obtained using the information stored in the unique name database 103. Obtaining information stored in the company unique name database 105, the product unique name database 106, and the company product unique name database 107 ・ Using the information stored in the company product database 104, the company product unique name De Acquiring information stored in the database 107 ・ Using information stored in the company unique name database 105 to acquire information stored in the company product unique name database 107 ・ Product unique name database 106 Information stored in the company product unique name database 107 using the information stored in. To get

  The knowledge proliferation rule is a rule for acquiring the remaining two or one type of information using one or two types of information from a company name, a product name, or a unique name of a product from a document. Specifically, the first knowledge proliferation rule, the second knowledge proliferation rule, the third knowledge proliferation rule, the fourth knowledge proliferation rule, the fifth knowledge proliferation rule, the sixth knowledge proliferation rule, and the seventh knowledge There are a proliferation rule, an eighth knowledge proliferation rule, a ninth knowledge proliferation rule, a tenth knowledge proliferation rule, an eleventh knowledge proliferation rule, and a twelfth knowledge proliferation rule.

  The first knowledge proliferation rule acquires a combination of a company name and a product name from a document using the company name. Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The first knowledge proliferation rule may be, for example, a sentence pattern in which a product name related to a company name can be directly extracted from a document in which the company name is described. It may be a sentence pattern that identifies the existing sentence. If the document is a Web page, it may be an HTML tag structure pattern that directly extracts a product name related to the company name from the Web page on which the company name is described. It may be an HTML tag structure pattern that identifies a sentence that is being read. Moreover, the same pattern as the above which cannot extract the product name related to the company name from the document in which the company name is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of the company name described in the “% company name%” part and the product name described in the “% product name%” part, the “<title>% company name% (Handling)? Product </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A. *?>% Product name% </a> </ Li>) + </ ul> ".

  The second knowledge proliferation rule uses the company name from the document to obtain a combination of the company name and the unique name of the product

  Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The second knowledge proliferation rule may be, for example, a sentence pattern in which a proper noun of a product related to a company name can be directly extracted from a document in which the company name is described. It may be a sentence pattern that identifies sentences in which nouns are written. When the document is a Web page, it may be an HTML tag structure pattern that directly extracts a proper noun of a product related to the company name from the Web page in which the company name is described. The pattern of the HTML tag structure which specifies the sentence in which the noun is described may be sufficient. Moreover, the same pattern as the above which cannot extract the proper noun of the product related to the company name from the document in which the company name is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of the company name described in the “% company name%” part and the product unique name described in the “% unique name%” part, the “<title>% company name” % (Handling)? Product </ title>. *? <Span. *?> Product list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% < / A> </ li>) + </ ul> ".

  The third knowledge proliferation rule obtains a set of a company name, a product name, and a product unique name from a document using the company name.

  Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The third knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a product name related to a company name and a proper noun of the product among documents in which the company name is described, It may be a sentence pattern that directly extracts the product name and identifies a sentence that includes the unique name of the product, or a sentence pattern that identifies a sentence that includes the product name and directly extracts the proper noun of the product. And a sentence pattern that identifies a sentence in which the proper noun of the product is described. If the document is a web page, it may be a pattern of an HTML tag structure that directly extracts a product name related to the company name and a proper noun of the product from the web page in which the company name is described, It may be an HTML tag structure pattern that directly identifies the product name and identifies a sentence that includes the unique name of the product, or an HTML tag structure pattern that identifies the sentence including the product name and directly identifies the proper noun of the commodity. Alternatively, it may be an HTML tag structure pattern that specifies a sentence in which a product name and a proper noun of the product are described. Moreover, the same pattern as the above in which the product name related to the company name and the proper noun of the product cannot be extracted from the document in which the company name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The fourth knowledge proliferation rule acquires a combination of a company name and a product name from a document using a product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The fourth knowledge proliferation rule may be, for example, a sentence pattern in which a company name related to a product name can be directly extracted from a document in which the product name is described, and the company name is described. It may be a sentence pattern that identifies the existing sentence. If the document is a Web page, it may be an HTML tag structure pattern that directly extracts the company name related to the product name from the Web page on which the product name is described. The pattern of the HTML tag structure which identifies the sentence currently carried out may be sufficient. In addition, a pattern similar to the above in which a company name related to a product name cannot be extracted from documents in which the product name is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of the company name described in the “% company name%” part and the product name described in the “% product name%” part, the “<title>% company name% (Handling)? Product </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A. *?>% Product name% </a> </ Li>) + </ ul> ".

  The fifth knowledge proliferation rule acquires a set of a product name and a unique name of the product from the document using the product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The fifth knowledge proliferation rule may be, for example, a sentence pattern in which a unique name of a product related to the product name can be directly extracted from a document in which the product name is described. It may be a sentence pattern that identifies a sentence in which a name is described. If the document is a Web page, it may be an HTML tag structure pattern that directly extracts the unique name of the product related to the product name from the Web page describing the product name. It may be an HTML tag structure pattern that identifies a sentence in which a unique name is described. Moreover, the same pattern as the above which cannot extract the unique name of the product related to the product name among the documents in which the product name is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of a product name described in the “% product name%” part and a unique name of the product described in the “% unique name%” part, “<span. *?> % Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> .

  The sixth knowledge proliferation rule acquires a combination of a company name, a product name, and a product unique name from a document using the product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The sixth knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a company name related to a product name and a proper noun of the product among documents in which the product name is described, It may be a sentence pattern that directly extracts the company name and identifies the sentence including the unique name of the product, or may be a sentence pattern that identifies the sentence including the company name and directly extracts the proper noun of the product. And a sentence pattern that identifies a sentence in which the proper noun of the product is described. If the document is a Web page, the HTML tag structure pattern may be used to directly extract the company name and product proper nouns that are related to the product name from the Web page on which the product name is described. The HTML tag structure pattern that directly identifies the company name and identifies the sentence including the unique name of the product may be used, or the HTML tag structure that identifies the sentence including the company name and directly identifies the proper noun of the commodity. A pattern may be sufficient, and the pattern of the tag HTML tag structure which specifies the text in which the company name and the proper noun of the product are described may be used. Moreover, the same pattern as the above in which the company name related to the product name and the proper noun of the product cannot be extracted from the document in which the product name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The seventh knowledge proliferation rule acquires a combination of a company name and a product unique name from a document using the product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The seventh knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a company name related to a unique name of a product from documents in which the unique name of the product is described. It may be a sentence pattern that identifies sentences in which names are described. If the document is a Web page, it may be a pattern of an HTML tag structure that directly extracts a company name related to the unique name of the product from the Web page in which the unique name of the product is described. The pattern of the HTML tag structure which specifies the text in which the company name is described may be sufficient. Moreover, the same pattern as the above which cannot extract the company name related to the unique name of the product among the documents in which the unique name of the product is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of the company name described in the “% company name%” part and the product unique name described in the “% unique name%” part, the “<title>% company name” % (Handling)? Product </ title>. *? <Span. *?> Product list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% < / A> </ li>) + </ ul> ".

  The eighth knowledge proliferation rule acquires a set of a product name and a product unique name from a document using the product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The eighth knowledge proliferation rule may be, for example, a sentence pattern that can directly extract the product name related to the product unique name from the document in which the product unique name is described. It may be a sentence pattern that specifies a sentence in which a name is described. Further, when the document is a Web page, it may be an HTML tag structure pattern that directly extracts the product name related to the product unique name from the Web page in which the product unique name is described. It may be an HTML tag structure pattern that identifies a sentence in which a product name is described. Moreover, the same pattern as the above which cannot extract the product name relevant to the product unique name among the documents in which the product unique name is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, when acquiring a combination of a product name described in the “% product name%” part and a unique name of the product described in the “% unique name%” part, “<span. *?> % Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> .

  The ninth knowledge proliferation rule acquires a set of a company name, a product name, and a product unique name from a document using the product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The ninth knowledge proliferation rule is, for example, a sentence pattern in which a company name and a product name related to a product unique name can be directly extracted from a document in which a product unique name is described. It may be a sentence pattern that directly extracts a company name and identifies a sentence including the product name, or a sentence pattern that identifies a sentence including the company name and directly extracts the product name. It may be a sentence pattern that specifies a sentence in which a name is described. If the document is a Web page, the HTML tag structure pattern that directly extracts the company name and the product name related to the product unique name from the Web page on which the product unique name is described. It may be an HTML tag structure pattern that directly identifies a company name and identifies a sentence including a product name, or an HTML tag structure pattern that identifies a sentence including a company name and directly identifies a product name. It may be a pattern of an HTML tag structure that specifies a sentence in which a company name and a product name are described. Moreover, the same pattern as the above which cannot extract the company name and the product name related to the unique name of the product among the documents in which the unique name of the product is described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The tenth knowledge proliferation rule acquires a combination of a company name, a product name, and a unique name of a product from a document using a combination of the company name and the product name.

  Using a combination of a company name and a product name includes, for example, one or more company names stored in the company database 101 and one or more product names stored in the product database 102 from the document described above. It is also possible to select a document that has a name of a company and a character string around the product name.

  The tenth knowledge proliferation rule is, for example, a sentence pattern in which a unique name of a product related to a company name and a product name can be directly extracted from documents in which the company name and the product name are described. Alternatively, it may be a sentence pattern that specifies a sentence in which the unique name of the product is described. Further, when the document is a Web page, it is an HTML tag structure pattern that directly extracts a unique name of a product related to the company name and the product name from the Web page describing the company name and the product name. It may be a pattern of an HTML tag structure that specifies a sentence in which a proper name of a product is described. Moreover, the same pattern as the above which cannot extract the unique name of the product related to the company name and the product name among the documents in which the company name and the product name are described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The eleventh knowledge proliferation rule acquires a combination of a company name, a product name, and a unique name from a document using a combination of the company name and the unique name of the product.

  Using a combination of company name and unique name means, for example, one or more company names stored in the company database 101 from the above-mentioned document and one or more unique names of one or more products stored in the unique name database 103 May be selected, or a peripheral character string in which the company name and the unique name of the product are described may be extracted.

  The eleventh knowledge proliferation rule is, for example, a sentence pattern in which a product name related to a company name and a product unique name can be directly extracted from a document in which the company name and product unique name are described. It may be a sentence pattern that identifies a sentence in which a product name is described. If the document is a Web page, an HTML tag structure pattern that directly extracts the product name related to the company name and the product unique name from the Web page describing the company name and the product unique name. It may be a pattern of an HTML tag structure that specifies a sentence in which a product name is described. Moreover, the same pattern as the above which cannot extract the product name related to the company name and the unique name of the product among the documents in which the company name and the unique name of the product are described may be used. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The twelfth knowledge proliferation rule acquires a combination of a company name, a product name, and a unique name from a document using a combination of a product name and a unique name of the product.

  Using a combination of a product name and a unique name means, for example, one or more product names stored in the product database 102 from the above-mentioned document and a unique name of one or more products stored in the unique name database 103. May be selected, a product name and a surrounding character string in which a product unique name is described may be extracted.

  The twelfth knowledge proliferation rule is, for example, a sentence pattern that can directly extract a company name related to a product name and a product unique name from a document in which a product name and a product unique name are described. It may be a sentence pattern that specifies a sentence in which a company name is described. If the document is a Web page, the HTML tag structure pattern that directly extracts the company name related to the product name and the product unique name from the Web page describing the product name and the product unique name. It may be a pattern of an HTML tag structure that specifies a sentence in which a company name is described. Moreover, the same pattern as the above in which the company name related to the product name and the product unique name cannot be extracted from the document in which the product name and the product unique name are described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the above-described rules may be combined. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. When acquiring a set of unique names of “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. * ?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> ”.

  The knowledge proliferation rule database 111 may store not only the knowledge proliferation rules but also the types of knowledge proliferation rules, the number of acquisitions of knowledge proliferation rules, the acquisition ratio, or the objects extracted by the knowledge proliferation rules. Targets to be extracted by knowledge proliferation rules are the cases where a name such as a company name, a product name, or a unique name of a product is acquired as a result of using the knowledge proliferation rule from a document, and a sentence including a unique name of a product is acquired. If there is a possibility of acquiring multiple types of information, such as the case, it is a value that clearly indicates what to acquire. Specifically, a value such as “name” or “text” may be stored.

  The knowledge proliferation rule database 111 may hold each of the above knowledge proliferation rules as a flag, or may be a separate database for each of the above knowledge proliferation rules.

  The company name acquisition unit 112 acquires one or more company names from one or more Web server devices using one or more company rules, and accumulates them in the company database 101.

  Specifically, the Web server device is a server device that publishes a Web service or a database. The Web service is not limited to publishing what is configured in a markup language such as HTML format. For example, a system of application software operable on a Web browser or a dedicated application execution environment May be published. The Web server device may be published on the Internet or may be published on a local area.

  The company name acquisition unit 112 acquires a company name using a company rule from a Web page acquired from a Web server device. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  Acquiring a company name using a company rule means acquiring a company name from a company rule registered in the company rule database 108 from a document acquired from a Web server device or an HTML structure. For example, a character string that applies to a portion described as “% company name%” in an arbitrary company rule is acquired.

  When the company name is stored in the company database 101, a certain threshold may be set for the number of acquisitions and the acquisition ratio of the used company rules, and only information that exceeds the threshold may be registered in the company database 101. . The threshold value may be arbitrarily set according to the operation.

  The product name acquisition unit 113 acquires one or more product names from one or more Web server devices using one or more product rules, and accumulates them in the product database 102.

  The product name acquisition unit 113 acquires the product name from the Web page acquired from the Web server device using the product rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  Acquiring a product name using a product rule is to acquire a product name from a document acquired from a Web server device, an HTML structure, or the like. For example, a character string corresponding to a portion described as “% product name%” in an arbitrary product rule is acquired.

  When accumulating product names in the product database 102, a certain threshold may be set for the number of times and the acquisition ratio of the used product rules, and only information that exceeds the threshold may be registered in the product database 102. . The threshold value may be arbitrarily set according to the operation.

  The unique name acquisition unit 114 acquires one or more unique names from one or more Web server devices using one or more unique name rules, and accumulates them in the unique name database 103.

  The unique name acquisition unit 114 acquires the unique name of the product from the Web page acquired from the Web server device using the unique name rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  Acquiring the unique name of the product using the unique name rule is to acquire the unique name of the product from the document acquired from the Web server device or the structure of HTML. For example, a character string corresponding to a portion described as “% unique name%” in an arbitrary unique name rule is acquired.

  When storing the unique name of the product in the unique name database 103, a certain threshold is set for the number of acquisitions and the acquisition ratio of the used unique name rule, and only information that exceeds the threshold is registered in the unique name database 103. You may do it. The threshold value may be arbitrarily set according to the operation.

  The knowledge acquisition unit 115 uses one or more company names and one or more knowledge proliferation rules in the company database 101 from one or more Web server devices, and one or more company names and product names, or one or more. The company name and unique name, or one or more company names, product names, and unique names are acquired, and the acquired information is stored in the corresponding database. Alternatively, using one or more product names and one or more knowledge proliferation rules in the product database 102, one or more company names and product names, or one or more product names and unique names, or one or more companies The name, product name, and unique name are acquired, and the acquired information is stored in the corresponding database. Alternatively, using one or more unique names and one or more knowledge proliferation rules in the unique name database 103, one or more company names and unique names, or one or more product names and unique names, or one or more unique names The company name, product name, and unique name are acquired, and the acquired information is stored in the corresponding database. Alternatively, one or more company names, product names, and unique names are acquired using one or more pairs of one or more company names and product names in the company product database 104 and one or more knowledge proliferation rules. Are stored in the company product unique name database 107. Alternatively, one or more company names, product names, and unique names are acquired using one or more pairs of one or more company names and product unique names in the company unique name database 105 and one or more knowledge proliferation rules. The acquired information is stored in the company product unique name database 107. Alternatively, one or more company names, product names, and unique names are acquired using one or more pairs of one or more product names and product unique names in the product unique name database 106 and one or more knowledge proliferation rules. The acquired information is stored in the company product unique name database 107.

  The knowledge acquisition unit 115 acquires a company name, a product name, or a unique name of a product from the Web page acquired from the Web server device using the knowledge multiplication rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  One or more company names and product names, or one or more company names and unique names, or one or more company names, using one or more company names and one or more knowledge proliferation rules in the company database 101 The acquisition of the product name and the unique name means that the company name stored in the company database 101 appears from the document acquired from the Web server device or the HTML structure, and is stored in the knowledge proliferation database. It is to acquire two or more sets of a company name, a product name, or a unique name of a product using the one knowledge growth rule, the second knowledge growth rule, or the third knowledge growth rule. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  One or more company names and product names, or one or more product names and unique names, or one or more company names using one or more product names and one or more knowledge proliferation rules in the product database 102 The acquisition of the product name and the unique name means that the product name stored in the product database 102 appears from the document acquired from the Web server device or the HTML structure, and is stored in the knowledge proliferation database. Using the fourth knowledge proliferation rule, the fifth knowledge proliferation rule, or the sixth knowledge proliferation rule, obtaining two or more sets of a company name, a product name, or a unique name of a product. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  One or more company names and unique names, or one or more product names and unique names, or one or more company names using one or more unique names and one or more knowledge proliferation rules in the unique name database 103. The product name and unique name are acquired from the document acquired from the Web server device or the structure of HTML, etc., and the unique name of the product stored in the unique name database 103 appears and is stored in the knowledge proliferation database. Using the seventh knowledge proliferation rule, the eighth knowledge proliferation rule, or the ninth knowledge proliferation rule, two or more sets of company names, product names, or product unique names are obtained. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  The acquisition of one or more company names, product names, and unique names using one or more pairs of one or more company names and product names in the company product database 104 and one or more knowledge proliferation rules is performed from the Web server device. From the acquired document or HTML structure, the company name in the company database 101 and the product name stored in the product database 102 appear, and using the tenth knowledge proliferation rule accumulated in the knowledge proliferation database, It is to acquire a set of company name, product name, and product unique name. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  The acquisition of one or more company names, product names, and unique names using one or more pairs of one or more company names and product unique names in the company unique name database 105 and one or more knowledge proliferation rules From the document acquired from the server device or the structure of HTML, the company name of the company database 101 and the unique name of the product stored in the unique name database 103 appear and the eleventh stored in the knowledge proliferation database It is to acquire a set of a company name, a product name, and a unique name of a product using the knowledge proliferation rule. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  The acquisition of one or more company names, product names, and unique names using one or more pairs of one or more product names and product unique names in the product unique name database 106 and one or more knowledge proliferation rules The product name stored in the product database 102 and the unique name of the product registered in the unique name database 103 appear from the document acquired from the server device or the HTML structure, and are stored in the knowledge proliferation database. The combination of the company name, the product name, and the unique name of the product is acquired using the twelfth knowledge proliferation rule. For example, the character string that applies to the part described as “% company name%” in each rule is the company name, the character string that applies to the part described as “% product name%” is the product name, and “% unique name%” "Is acquired as the unique name of the product.

  When accumulating company names, product names, and product unique names in the corresponding databases, only information that exceeds the threshold by setting a certain threshold for the number of acquisitions and acquisition ratio of the knowledge proliferation rules used May be registered in each database. The threshold value may be arbitrarily set according to the operation.

  The rule learning unit 116 searches for one or more Web server devices using one or more pieces of information existing in one or more databases, acquires one or more patterns in which one or more pieces of information appear, and the patterns are stored in advance. One or more patterns that appear more frequently as the predetermined condition is satisfied are acquired, and the one or more patterns are accumulated in the corresponding rule databases.

  Specifically, searching for one or more Web server devices using one or more information existing in each of one or more databases means company database 101, product database 102, unique name database 103, or company product. Searching a server device using one or more arbitrary information stored in one or more databases from the database 104, the company unique name database 105, or the product unique name database 106 as a search keyword.

  Searching for a Web server device is, for example, searching for character string information included in a Web page provided by a Web server device (not shown) on the Internet. Web search is performed using a search site, for example. Here, for example, transmitting a query including a search key to a so-called search engine provided by a search site and receiving information indicating a search result transmitted by the search engine is considered to be a web search, for example. . The search key can be transmitted by using an API provided by the search engine.

  To acquire one or more patterns in which one or more information appears is a pattern including one or more information existing in one or more databases used for the search from information acquired as a result of searching the server device. Is to get. The pattern may be, for example, a sentence pattern. If the target is a Web page, the pattern may be an HTML tag structure pattern, or the above-described rules may be combined. The acquisition of the pattern may be anything that can acquire the company name, the product name, and the unique name of the product, for example, the entire sentence may be acquired as a pattern, and the longest including information used for the search from multiple sentences The common partial sequence may be acquired.

To acquire one or more patterns that appear more frequently when a predetermined condition is satisfied, a condition to be registered in each rule database is determined in advance, and only patterns that satisfy the condition are acquired.
The predetermined condition is a condition that is registered in advance in each rule database. The condition may be any value that can evaluate the pattern. For example, it may be the number of times the pattern is acquired, or may be the ratio occupied by the pattern among all the acquired patterns.

  To store one or more patterns in each corresponding rule database, a pattern that appears better enough to satisfy a predetermined condition is registered in the database corresponding to the pattern. The corresponding database is a company rule database 108 for a pattern for acquiring a company name, a product rule database 109 for a pattern for acquiring a product name, and a unique name rule database 110 for a pattern for acquiring a unique name of a product. A pattern for acquiring a pair of company name and product name, a pattern for acquiring a pair of company name and product unique name, a pattern for acquiring a pair of product name and product unique name, and a company name and product name If it is a pattern for acquiring a set of unique names of products, it is registered in the knowledge proliferation rule database 111.

  In addition, when the rule learning unit 116 acquires one or more patterns using one or more information existing in a database storing two or more types of information sets, the rule learning unit 116 sets the two or more types of information sets. One or more patterns are stored in one or more rule databases corresponding to two or more types of information sets using the corresponding rule database and a part of information of two or more types of information sets.

  The set of two or more types of information is information on a company name and a product name, or a company name and a product unique name, or a product name and a product unique name, or a company name, a product name, and a product unique name. The database storing a set of two or more types of information is the company product database 104, the unique name database 103, the product unique name database 106, or the company product unique name database 107. The case where one or more patterns are acquired is a case where a pattern including two or more combinations among a company name, a product name, and a unique name of a product is acquired using a combination of two or more types of information. Using a part of information of two or more types of information means that each of the information is a combination of two types of information, each of which is a set of three types of information, and any two types All combinations of information sets are used. A pattern is acquired using the decomposed information, and the acquired pattern is registered in the corresponding database. For example, when the combination of two or more types of information is a company name and a product name, not only a pattern for acquiring a company name and a product name but also a pattern for acquiring a company name and a pattern for acquiring a product name Are also stored in the corresponding knowledge growth rule database 111, company rule database 108, or product rule database 109.

  FIG. 2 is a block diagram of the database construction device 2 in the present embodiment. The database construction device 2 includes a company product unique name database 107, a product similar group code correspondence database 21, a company similar group code unique name database 22, and a company similar group code unique name database construction unit 23.

  The product similar group code correspondence database 21 is a database that stores product names and similar group codes in association with each other. The product name is the name of a product or service that is sold or provided by a corporation and an individual, like the product name stored in the company product unique name database 107. The names of products and services here do not necessarily conform to the names of products and services described in the similar group code table. The similar group code is a code assigned to a product or service that is considered to be similar to each other in accordance with trademark examination standards, and is a code of 2 digits, 1 alphabet, and 2 digits. (Http://www.jpo.go.jp/shiryou/kijun/kijun2/ruiji_kijun9.htm)

  The product names and the similar group codes in the product similar group code correspondence database 21 are normally stored in a one-to-many relationship, and the same similar group codes are stored in association with a plurality of product names.

  The company similar group code unique name database construction unit 23 obtains a product name from a set of one or more company names, product names, and product unique names stored in the company product unique name database 107, and obtains the obtained product name. The similar group code of the product is acquired from the product similar group code correspondence database 21, and the combination of the company name, the acquired similar group code, and the unique name of the product is stored in the company similar group code unique name database 22.

  FIG. 3 is a block diagram of the trademark infringement detection apparatus 3 in the present embodiment. The trademark infringement detection device 3 includes a database 31, a reception unit 32, a trademark class determination unit 33, a product class determination unit 34, and an output unit 35.

  The database 31 may be the company product unique name database 107 constructed by the database construction device 1 or the company similar group code unique name database 22 constructed by the database construction device 2.

  The receiving unit 32 receives survey target trademark information having code specifying information that is information for specifying a similar group code of a trademark and a product. The investigation target trademark information is, for example, a set of a trademark character string that is a trademark character string and one or more code specifying information. The code specifying information is a character string that specifies a similar group code such as a product name or the similar group code itself. The trademark character string and the character string of the product name are composed of one or more characters. The characters constituting the trademark character string and the product name character string may be of any type, such as kanji, alphabet, katakana, and hiragana. Two or more types of characters may be mixed. Acceptance here means acceptance of information input from input devices such as a keyboard, mouse, touch panel, reception of information transmitted via a wired or wireless communication line, recording of an optical disk, magnetic disk, semiconductor memory, etc. It is a concept including reception of information read from a medium. Any means can be used for inputting the trademark information to be investigated, such as a numeric keypad, a keyboard, a mouse, or a menu screen. The receiving unit 32 can be realized by a device driver for input means such as a numeric keypad or a keyboard, control software for a menu screen, and the like.

  The trademark similarity determination unit 33 determines the similarity of a trademark using the unique names of one or more products included in the database 31 and the trademark included in the trademark information to be investigated.

  FIG. 4 is a block diagram of the trademark similarity determination unit 33 of the trademark infringement detection device 3 according to the present embodiment. The trademark similarity determination unit 33 includes a trademark character string acquisition unit 331, a designation acquisition unit 332, a designation display unit 333, a designation designation reception unit 334, a similarity search unit 335, and a trademark similarity determination result acquisition unit 336.

  The trademark character string acquisition unit 331 receives a trademark character string that is a trademark character string. The trademark character string is composed of one or more characters. The characters constituting the trademark character string may be of any type, such as kanji, alphabet, katakana and hiragana.

  The name acquisition unit 332 acquires name information, which is information on the name of the trademark character string, from the trademark character string received by the trademark character string acquisition unit 331. The name acquisition unit 332 may acquire a plurality of pieces of name information from one trademark character string. The name information acquired by the name acquisition unit 332 is information such as hiragana, katakana, or the like. Which of these names is determined by, for example, the specification of a similarity search unit 335 (described later) that uses the name information for searching.

  The name acquisition unit 332 may acquire the name information in any way. For example, the name acquisition unit 332 uses a dictionary having one or more character strings and information on reading of the character strings in association with each other, and uses a trademark character string stored in a storage medium (not shown). One or more pieces of name information corresponding to the trademark character string received by the acquisition unit 331 may be automatically acquired. Further, one or more pieces of name information corresponding to a trademark character string input from a user or the like via an input device (not shown) may be acquired (accepted). In this embodiment, a case where name information is automatically acquired will be described as an example.

  For example, the name acquisition unit 332 performs morphological analysis on the trademark character string, and acquires the reading information of the trademark character string obtained by the morphological analysis as name information. For example, by using morphological analysis, it is possible to decompose a trademark character string into morphemes, acquire readings of each decomposed character string, and acquire the designation information of the trademark character string by combining these character strings. is there. In morpheme analysis, morpheme decomposition and acquisition of readings from character strings decomposed into morphemes are performed using, for example, a dictionary that associates character strings that can be morphemes with information about the readings, as described above. Done. As the morphological analysis, for example, a morphological analysis system such as “Mecab” (http://mecab.sourceforge.net/) or “ChaSen” (http://chasen.naist.jp) Etc. are available.

  The name acquisition unit 332 may acquire name information by converting a kanji, a number, a character such as an alphabet, or a word constituting a trademark character string into a reading. The process of converting kanji, numeric characters, and words into readings can be realized by, for example, a so-called re-conversion technique of a Japanese input method or a morphological analysis system technique as described above. Further, the process of converting alphabetic characters and words into readings can be realized, for example, by a technique for converting the input of the alphabet string of the Japanese input method into Japanese. A technique for converting alphabet strings into Japanese readings is disclosed in, for example, Japanese Patent Application Laid-Open No. 2009-199434. In addition, since an alphabet, a kanji, and a numerical name (reading) of a number are not necessarily one way, the name acquisition part 332 may acquire several information of various combinations. For example, a plurality of designations “IT”, “IT” and “It” may be acquired from “IT”. A plurality of designations “Yamada”, “Yamata”, “Yamaden”, “Sanden”, “Sanda”, and “Santa” may be acquired from “Yamada”.

  Further, the name acquisition unit 332 accumulates in advance a character string (for example, kanji, alphabet, numbers, etc.) that can constitute a trademark character string and reading information (dictionary information) in a storage unit (not shown). In this way, one or more reading information may be acquired for each character constituting the trademark character string, and a combination of the acquired reading information may be acquired as name information.

  Further, for example, with respect to morpheme dictionary information prepared in advance having one or more character strings associated with katakana or hiragana indicating the reading, the name acquisition unit 332 may include other than katakana or hiragana included in the trademark character string. This can be realized by searching for a character string that matches the character string and sequentially acquiring reading (name) information associated with the detected character string from the dictionary information. When there are multiple readings for a single character string, or when different designations are obtained by dividing the number of characters used in the search, a combination of multiple readings or changing the dividing may be used to appropriately combine a plurality of readings. Get name information. Such dictionary information may be held in advance by the name acquisition unit 332, for example. The name acquisition unit 332 may include a storage medium that holds such dictionary information.

  The name acquisition unit 332 may further acquire one or more divided name information by dividing the one or more name information acquired above. Specifically, the division name information is information of a part of the name information. The divided name information is used as substantially the same as the name information in, for example, a similarity search described later. For example, the name acquisition unit 332 divides each character string of the acquired name information at an arbitrary position and acquires character strings before or after being divided by the divided position, or both as divided name information. To do. The name acquisition unit 332 may sequentially change the position where the character string is divided, and acquire the divided name information according to each position where the character string is divided. For example, the name acquisition unit 332 may acquire all the character strings before and after the delimiter positions obtained when the character strings constituting the name information are divided between the characters as the divided name information. .

  The name display unit 333 displays one or more pieces of name information acquired by the name acquisition unit 332 on a display device such as a monitor (not shown). The name display unit 333 may display one or more pieces of divided name information acquired by the name acquisition unit 332. The name display unit 333 may be considered as including or not including a display device. The designation display unit 333 can be realized by display device driver software, or display device driver software and a display device.

  The designation designation accepting unit 334 accepts designation designation information for designating one or more pieces of designation information acquired by the designation acquisition unit 332 from a user or the like. For example, the designation designation information accepts designation designation information for designating one or more pieces of designation information displayed on the designation display unit 333. Further, the designation designation accepting unit 334 may accept designation designation information for designating one or more of the designation information and the divided designation information acquired by the designation acquisition unit 332 from a user or the like. For example, the designation designation accepting unit 334 accepts designation designation information for designating one or more of the designation information and the divided designation information displayed by the designation display unit 333. Incidentally, the nominal designation receiver 334, by providing a rule that defines how to specify in advance what nominal, may be may not be selected by the user. The name designation information input means may be anything such as a numeric keypad, keyboard, mouse, or menu screen. The designation designation accepting unit 334 can be realized by a device driver for input means such as a numeric keypad or a keyboard, control software for a menu screen, or the like.

  The similarity search unit 335 performs a similarity search using each name information acquired by the name acquisition unit 332. Alternatively, the similarity search unit 335 may perform a similar search using each piece of name information specified by the name designation information among the name information acquired by the name acquisition unit 332. Further, the similarity search unit 335 may perform a similarity search using each piece of nominal information acquired by the nominal acquisition unit 332 and each piece of divided nominal information. Alternatively, the similarity search unit 335 performs a similar search by using the name information and the divided name information specified by the name designation information among the name information and the divided name information acquired by the name acquisition unit 332. May be. The use of the designation name information and the division designation information specified by the designation designation information means, for example, that only the designated designation information or division designation information is used. Similarity search unit 335 is similar to name information (or divided name information) from registered trademark name information (hereinafter referred to as registered name information) stored in a storage unit such as a storage medium (not shown). Search registered name information. The registered name information is created and constructed in advance from the unique names in the company product unique name database 107 or the company similar group code unique name database 22.

  Note that a storage unit (not shown) in which registered trademark designation information is stored may be included in the trademark kind determination unit 33 or an external server device (not illustrated). In addition, the similar search unit 335 transmits the name information or the divided name information and the similar group code to a server device or the like that performs a similar search process for a registered trademark (not shown), and the similar search unit 335 receives the search result. In such a case, it may be considered that the similarity search unit 335 has performed a similarity search.

  Further, the similarity search process by the similarity search unit 335 may be an acoustic determination process of name information. Detects a factor that is judged to be dissimilar from syllables of nominal information, converts the factor into a penalty value, determines whether the nominal information is similar using the converted penalty value, and uses a trademark character string This is a process of detecting the name information similar to the name information acquired from the name information of the registered trademark. Here, for convenience of explanation, the division name information is also referred to as name information. In addition, here, it is assumed that registered name information, which is name information of a registered trademark, is stored in a storage unit (not shown) in association with similar group code and registered trademark identification information.

  The result of the similarity search acquired by the similarity search unit 335 is information that can indicate, for example, whether or not at least similar registered name information is detected. For example, the result of the similarity search may be information indicating the presence or absence of similar registered name information, or may be information indicating one or more similar registered name information. Further, it may be information indicating the number of similar registered name information. In addition, when the name information is stored in the above-described storage unit (not shown) in association with a registered trademark or its identification information (for example, a registration number), the registered trademark corresponding to similar registered name information or its identification Information may be acquired from a storage unit (not shown) and output as a search result. Moreover, you may output the information containing the information of the similar group code matched with registration name call information. Specific processing for performing a similar search for a registered trademark name is known as a technique such as the above-described Patent Document 1, a name search service provided by the Japanese Patent Office, or a name search service for a commercial database. Therefore, detailed description is omitted here.

  The trademark similarity determination result acquisition unit 336 acquires, for example, the result of the similarity search by the similarity search unit 335, the company name and the similar group code specifying information related to the name information of the result of the similarity search. As for the result of the similarity search, the result of the similarity search is acquired for each name information for which the similarity search is performed. As for the result of the similarity search, for example, the result of the similarity search is acquired for each divided name information that has been subjected to the similarity search. In addition, a registered trademark determined to be similar in name and information indicating identification information of the registered trademark may be added to each piece of evidence information.

  The merchandise similarity determination unit 34 determines the similarity of the merchandise using one or more product names or one or more similar group codes included in the database 31 and code identification information included in the survey target trademark information. When the survey target trademark information holds the product name, the company name and the product unique name of the record having the product name related to the product name of the survey target trademark information are acquired from the company product unique name database 107. The term “related to the product name of the survey target trademark information” means, for example, the case where the product name of the survey target trademark information and the product name of the company product unique name database 107 are the same or similar in meaning. is there. When the survey target trademark information holds a similar group code, the company name and product unique name of the record having the same similar group code as the similar group code of the survey target trademark information are stored from the company similar group code specific name database 22. get.

  The output unit 35 outputs information on the similarity of the trademark and the similarity of the product based on the determination result of the trademark similarity determination unit 33 and the determination result of the product similarity determination unit 34. For example, information that can indicate whether or not similar information is detected is output. For example, in the case of the determination result of the trademark kind determination unit 33, information indicating the presence or absence of similar registered name information may be used, or information indicating one or more similar registered name information may be used. Moreover, the information which shows the number of similar registration name call information etc. may be sufficient. Further, it may be an arbitrary number of similar registered name information. Further, registered trademarks corresponding to similar registered name information and identification information thereof may be output. Further, all or all of the above output contents may be output by adding a part or all of the company name, the product name, the similar group code, or the unique name of the product. Further, for example, in the case of the determination result of the product type rejection determination unit 34, information indicating the presence / absence of corresponding code specifying information may be used, or information indicating one or more corresponding code specifying information may be used. Moreover, the information which shows the number of applicable code specific information etc. may be sufficient. Moreover, about all the said output content, you may add and output one or both among a company name or a specific name.

  The output unit 35 may output all the judgment results of the trademark kind judgment unit 33 and all judgment results of the product kind judgment unit 34, or may be either one. Further, the output unit 35 may output only the common product unique names among the results of the trademark similarity determination unit 33 and the product similarity determination unit 34.

  The output described here is a concept including display on a display, printing on paper or the like by a printer, transmission to an external device, accumulation in a recording medium, and the like. The output unit 35 may or may not include a display device. The output unit 35 can be realized by driver software of an output device or driver software of an output device and an output device.

  FIG. 5 is a flowchart showing an example of the operation of the company name acquisition unit 112 according to the present embodiment. The operation will be described below with reference to FIG.

  (Step S401) The company name acquisition unit 112 acquires a company rule from the company rule database 108. If it can be acquired, the process proceeds to step S402. If it cannot be acquired, the process ends.

  (Step S402) The company name acquisition unit 112 substitutes 1 for a counter m.

  (Step S403) The company name acquisition unit 112 determines whether or not there is an mth company rule in the company rules acquired in step S401. If there is, the process proceeds to step S404, and if not, the process ends.

  (Step S404) The company name acquisition unit 112 acquires a Web page from the Web server device. Get it at this time. The Web page may be a Web page that may contain a company name acquired in advance using a company rule that acquires a sentence including the company name from the company rule database 108, or is acquired in large quantities at random. You may do it. If the Web page can be acquired, the process proceeds to step S405. If the Web page cannot be acquired, the process proceeds to step S410.

  (Step S405) The company name acquisition unit 112 substitutes 1 for a counter n.

  (Step S406) The company name acquisition unit 112 determines whether there is an nth Web page in the Web page acquired in Step S404. If there is, the process proceeds to step S407, and if not, the process proceeds to step S410.

  (Step S407) The company name acquisition unit 112 acquires a company name for the nth Web page of the Web page acquired in Step S404 by using the mth company rule of the company rule acquired in Step S401. If it can be acquired, the process proceeds to step S408. If it cannot be acquired, the process proceeds to step S409.

  (Step S409) The company name acquisition unit 112 increments the counter n. Then, the process returns to step S406.

  (Step S410) The company name acquisition unit 112 increments the counter m. Then, the process returns to step S403.

  Hereinafter, a specific operation of the company name acquisition unit 112 in the present embodiment will be described. Note that the acquisition result of the company name acquisition unit 112 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The company name acquisition unit 112 acquires all rules for acquiring a company name from the company rule database 108. If the company rule database 108 holds data related to the target extracted by the company rule, only the company rule whose “company name” is the target is acquired. In addition, if values such as the number of acquisitions and acquisition ratio of company rules are held, only company rules exceeding an arbitrary threshold may be acquired. In this specific example, it is assumed that five types of company rules have been acquired.

  FIG. 6 is a diagram illustrating an example of data stored in the company rule database 108. The company rule database 108 has a pattern for acquiring a company name. As a specific example, company rules for acquiring a noun phrase corresponding to a part described as “% company name%” as a company name are listed.

  The company name acquisition unit 112 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The company name acquisition unit 112 acquires company names from all acquired Web pages using all acquired company rules. In this case, since there are 5 types of company rules and 100 types of Web pages, the company name is acquired 500 times. Specifically, if the company rule is “<title>% company name% corporation </ title>”, if “<title> sample trading corporation </ title>” is described in the acquired Web page. , Get “sample trading” part. However, nothing is acquired unless the content corresponding to the pattern is described.

  The company name acquisition unit 112 registers the company name in the company database 101 when the company name can be acquired. In this case, “Sample Trading” is registered. In addition, if the company rule database 108 holds values such as the number of acquisitions and the acquisition ratio of company rules, it may be registered together with the values, and the “<title> sample trading” around the acquired information. </ Title> "may be registered together.

  As described above, according to the present embodiment, a company name can be acquired from an arbitrary Web page.

  FIG. 7 is a flowchart showing an example of the operation of the product name acquisition unit 113 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S501) The product name acquisition unit 113 acquires a product rule from the product rule database 109. If it can be acquired, the process proceeds to step S502. If it cannot be acquired, the process ends.

  (Step S502) The product name acquisition unit 113 substitutes 1 for the counter m.

  (Step S503) The product name acquisition unit 113 determines whether or not the product rule acquired in step S501 includes the mth product rule. If there is, the process proceeds to step S504, and if not, the process ends.

  (Step S504) The product name acquisition unit 113 acquires a Web page from the Web server device. Get it at this time. The Web page may be a Web page that may contain a product name acquired in advance using a product rule that acquires a sentence including the product name from the product rule database 109, or may be acquired in large quantities at random. You may do it. If the Web page can be acquired, the process proceeds to step S505. If the Web page cannot be acquired, the process proceeds to step S510.

  (Step S505) The product name acquisition unit 113 assigns 1 to the counter n.

  (Step S506) The product name acquisition unit 113 determines whether the Web page acquired in Step S504 includes the nth Web page. If there is, the process proceeds to step S507, and if not, the process proceeds to step S510.

  (Step S507) The product name acquisition unit 113 acquires the product name for the nth Web page of the Web page acquired in Step S504, using the mth product rule of the product rule acquired in Step S501. If it can be acquired, the process proceeds to step S508. If it cannot be acquired, the process proceeds to step S509.

  (Step S509) The product name acquisition unit 113 increments the counter n. Then, the process returns to step S506.

  (Step S510) The product name acquisition unit 113 increments the counter m. Then, the process returns to step S503.

  Hereinafter, a specific operation of the product name acquisition unit 113 in the present embodiment will be described. Note that the acquisition result of the product name acquisition unit 113 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The product name acquisition unit 113 acquires all rules for acquiring product names from the product rule database 109. If the product rule database 109 holds data related to the object to be extracted by the product rule, only the product rule whose “product name” is the target is acquired. Further, if values such as the number of acquisitions of product rules and the acquisition ratio are held, only product rules that exceed an arbitrary threshold may be acquired. In this specific example, it is assumed that five types of product rules have been acquired.

  FIG. 8 is a diagram illustrating an example of data stored in the product rule database 109. The product rule database 109 has a pattern for acquiring product names. As a specific example, product rules for acquiring a noun phrase corresponding to a portion described as “% product name%” as a product name are listed.

  The product name acquisition unit 113 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The product name acquisition unit 113 acquires product names from all acquired Web pages using all acquired product rules. In this case, since there are 5 types of product rules and 100 types of Web pages, the product name is acquired 500 times. Specifically, the product rule is “<title> product name list </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?>% Product name% </ Li>) + </ ul> ”,“ <title> Product name list </ title>... (Omitted)... <Span> list </ span> <ul> <Li> <a href = ”tv. html "> TV </ li> <li> <a href =" dvd. html "> DVD player </ li> <li> <a href =" pc. If “html”> PC </ li> </ UL>... (omitted)..., “TV,” “DVD player,” “PC” is acquired. However, nothing is acquired unless the content corresponding to the pattern is described.

  The product name acquisition unit 113 registers the product name in the product database 102 when the product name can be acquired. In this case, “TV”, “DVD player”, and “PC” are registered. Also, if the product rule database 109 holds values such as the number of product rule acquisitions and the acquisition ratio, it may be registered together with the values, and the “<title> product name” around the acquired information. List </ title> ... (omitted) ... <span> list </ span> <ul> <li> <a href = ”tv. html "> TV </ li> <li> <a href =" dvd. html "> DVD player </ li> <li> <a href =" pc. html "> PC </ li> </ ul> ... (omitted) ..." may be registered together.

  As described above, according to the present embodiment, a product name can be acquired from an arbitrary Web page.

  FIG. 9 is a flowchart showing an example of the operation of the unique name acquisition unit 114 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 601) The unique name acquisition unit 114 acquires a unique name rule from the unique name rule database 110. If it can be acquired, the process proceeds to step S602. If it cannot be acquired, the process ends.

  (Step S602) The unique name acquisition unit 114 substitutes 1 for the counter m.

  (Step S603) The unique name acquisition unit 114 determines whether or not there is an mth unique name rule in the unique name rule acquired in step S601. If there is, the process proceeds to step S604, and if not, the process ends.

  (Step S604) The unique name acquisition unit 114 acquires a Web page from the Web server device. Get it at this time. The web page may be a web page that may include the unique name of the product acquired in advance using the unique name rule for acquiring the sentence including the unique name of the product from the unique name rule database 110. You may acquire a large amount at random. If the Web page can be acquired, the process proceeds to step S605. If the Web page cannot be acquired, the process proceeds to step S610.

  (Step S605) The unique name acquisition unit 114 substitutes 1 for the counter n.

  (Step S606) The unique name acquisition unit 114 determines whether or not the Web page acquired in Step S604 includes the nth Web page. If there is, the process proceeds to step S607, and if not, the process proceeds to step S610.

  (Step S607) The unique name acquisition unit 114 uses the mth unique name rule of the unique name rule acquired in step S601 for the nth Web page of the Web page acquired in step S604. get. If it can be acquired, the process proceeds to step S608. If it cannot be acquired, the process proceeds to step S609.

  (Step S609) The unique name acquisition unit 114 increments the counter n. Then, the process returns to step S606.

  (Step S610) The unique name acquisition unit 114 increments the counter m. Then, the process returns to step S603.

  Hereinafter, a specific operation of the unique name acquisition unit 114 in the present embodiment will be described. Note that the acquisition result of the unique name acquisition unit 114 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The unique name acquisition unit 114 acquires all the rules for acquiring the unique name of the product from the unique name rule database 110. If the unique name rule database 110 stores data related to the object extracted by the unique name rule, only the unique name rule for which “product unique name” is the target is acquired. In addition, if values such as the number of acquisitions and acquisition ratios of unique name rules are held, only unique name rules exceeding an arbitrary threshold value may be acquired. In this specific example, it is assumed that five types of unique name rules have been acquired.

  FIG. 10 is a diagram illustrating an example of data stored in the unique name rule database 110. The unique name rule database 110 has a pattern for acquiring a unique name of a product. As a specific example, specific name rules for acquiring a noun phrase corresponding to a portion described as “% proper name%” as a proper name of a product are listed.

  The unique name acquisition unit 114 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The unique name acquisition unit 114 acquires the unique name of the product from all the acquired web pages using all the acquired unique name rules. In this case, since the unique name rule is 5 types and the web page is 100 types, the unique name of the product is acquired 500 times. Specifically, the unique name rule is “<title> product list </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A. * ?>% Proper name% </a> </ li>) + </ ul> ”,“ <title> product list </ title> ... (omitted) ... <span > List </ span> <ul> <li> <a href = ”sc. html "> sample camera </ li> <li> <a href =" sc2. If “html”> sample camera 2 </ li> </ ul>... (omitted)..., “sample camera” and “sample camera 2” are acquired. However, nothing is acquired unless the content corresponding to the pattern is described.

  When the unique name of the product can be acquired, the unique name acquisition unit 114 registers it in the unique name database. In this case, “sample camera” and “sample camera 2” are registered. If the unique name rule database 110 holds values such as the number of acquisitions of the unique name rule and the acquisition ratio, it may be registered together with the values, and the “<title> product” around the acquired information may be registered. List </ title> ... (omitted) ... <span> list </ span> <ul> <li> <a href = ”sc. html "> sample camera </ li> <li> <a href =" sc2. html "> sample camera 2 </ li> </ ul> ... (omitted) ..." may be registered together.

  As described above, according to the present embodiment, a unique name of a product can be acquired from an arbitrary Web page.

  FIG. 11 is a flowchart showing an example of the operation of the knowledge acquisition unit 115 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 701) The knowledge acquisition unit 115 acquires a knowledge proliferation rule from the knowledge proliferation rule database 111. If it can be acquired, the process proceeds to step S702. If it cannot be acquired, the process ends.

  (Step S702) The knowledge acquisition unit 115 substitutes 1 for a counter m.

  (Step S703) The knowledge acquisition unit 115 determines whether or not the knowledge multiplication rule acquired in step S701 includes the mth knowledge multiplication rule. If there is, the process proceeds to step S704, and if not, the process ends.

  (Step S704) The knowledge acquisition unit 115 acquires information used for the m-th knowledge proliferation rule of the knowledge proliferation rule acquired in step S701 from each database. Specifically, the information used for the knowledge proliferation rule is obtained by acquiring the company name from the company database 101 in the case of the first to third knowledge proliferation rules, and in the case of the fourth to sixth knowledge proliferation rules. The product name is acquired from the product database 102, the unique name of the product is acquired from the unique name database 103 in the case of the seventh to ninth knowledge proliferation rules, and from the company database 101 and the product database 102 in the case of the tenth knowledge proliferation rule. The company name and the product name are acquired. In the case of the eleventh knowledge proliferation rule, the company name and the unique name of the product are acquired from the company database 101 and the unique name database 103. In the case of the twelfth knowledge proliferation rule, the product database 102 is acquired. The product name and the product unique name are acquired from the unique name database 103.

  (Step S705) The knowledge acquisition unit 115 substitutes 1 for a counter n.

  (Step S706) It is determined whether or not the information used for the knowledge proliferation rule acquired in step S704 includes the nth information. If there is, the process proceeds to step S707, and if not, the process proceeds to step S713.

  (Step S707) The knowledge acquisition unit 115 acquires a Web page from the Web server device. As the Web page acquired at this time, the Web page including the nth information among the information used for the knowledge proliferation rule acquired in step S704 is acquired. If the Web page can be acquired, the process proceeds to step S708. If the Web page cannot be acquired, the process proceeds to step S714.

  (Step S708) The knowledge acquisition unit 115 assigns 1 to the counter o.

  (Step S709) The knowledge acquisition unit 115 determines whether or not the web page acquired in step S707 includes the o-th web page. If there is, the process proceeds to step S710, and if not, the process proceeds to step S714.

  (Step S710) The knowledge acquisition unit 115 uses the m-th knowledge proliferation rule of the knowledge proliferation rule acquired in Step S701 for the o-th Web page of the Web page acquired in Step S707. , Or get the unique name of the product. If it can be acquired, the process proceeds to step S711. If it cannot be acquired, the process proceeds to step S712.

  (Step S711) The company name, product name, or product unique name acquired in step S710 is stored in the corresponding database.

  (Step S712) The knowledge acquisition unit 115 increments the counter o. Then, the process returns to step S709.

  (Step S713) The knowledge acquisition unit 115 increments the counter m. Then, the process returns to step S703.

  (Step S714) The knowledge acquisition unit 115 increments the counter n. Then, the process returns to step S706.

  Hereinafter, a specific operation of the knowledge acquisition unit 115 in the present embodiment will be described. Note that the acquisition result of the knowledge acquisition unit 115 shown in this specific example is prepared for convenience of explanation, and does not indicate actual data.

  The knowledge acquisition unit 115 acquires all rules for acquiring a company name, a product name, or a unique name of a product from the knowledge multiplication rule database 111. If the knowledge proliferation rule database 111 holds data related to the object extracted by the knowledge proliferation rule, only the knowledge proliferation rule whose “name” is the object is acquired. In addition, if values such as the number of acquisitions and the acquisition ratio of knowledge proliferation rules are held, only knowledge proliferation rules exceeding an arbitrary threshold may be acquired. In this specific example, it is assumed that five types of knowledge proliferation rules have been acquired.

  FIG. 12 is a diagram showing an example of data stored in the knowledge growth rule database 111. The knowledge proliferation rule database 111 has a pattern for acquiring a company name, a product name, or a unique name of a product. As a specific example, the noun phrase corresponding to the part described as “% company name%” is the company name, and similarly the noun phrase corresponding to the part described as “% product name%” is the product name, “% The knowledge proliferation rules for acquiring the noun phrases corresponding to the part described as “proprietary name%” as the proper name of the product are listed. In this specific example, instead of creating a database for each type of knowledge proliferation rule, the knowledge proliferation rule type is held in the knowledge proliferation database.

  The knowledge acquisition unit 115 acquires information to be used by the acquired knowledge proliferation rule. For example, when the first knowledge proliferation rule is acquired, all company names described in the company database 101 are acquired. Note that if the company database 101 holds values such as the number of acquisitions and the acquisition ratio of the company name, the company name below an arbitrary threshold value may not be acquired. In this specific example, it is assumed that ten types of company names have been acquired.

  The knowledge acquisition unit 115 acquires a web page. For example, in this specific example, assume that 100 types of Web pages are acquired for each acquired company name.

The knowledge acquisition unit 115 acquires company names from all acquired Web pages using all acquired knowledge proliferation rules. In this case, since there are 5 types of knowledge proliferation rules, 10 types of information used by the knowledge proliferation rules, and 100 types of Web pages, there are 1 or 2 types of information among 5000 times company name, product name, and product unique name. Acquire.
Specifically, the knowledge proliferation rule is “<title>% company name% (handling)? Product </ title>. *? <Span. *?> List </ span><ul.*?>(<Li .. *?><A.*?>% Product name% </a></li>) + </ ul> ”and the information used is“ sample industry ” , “<Title> sample industry products </ title>... (Omitted)... <Span> list </ span><ul><li><a href =” scr. html ”> screw </a></li><li><a href =“ driv. If “html”> driver </a></li></ul>... (omitted)..., “screw” and “driver” are acquired. However, nothing is acquired unless the content corresponding to the pattern is described.

  When new information can be acquired, the knowledge acquisition unit 115 registers it in each corresponding database. In this case, “screw” and “driver” are registered in the product database 102, “sample industry” and “screw”, and “sample industry” and “driver” are registered in the company product database 104. Note that if the knowledge proliferation rule database 111 holds values such as the number of times of knowledge proliferation rule acquisition and the acquisition ratio, it may be registered together with the values, and “<title>” is the periphery of the acquired information. Sample industry products </ title> ... (omitted) ... <span> list </ span> <ul> <li> <a href = ”scr. html ”> screw </a> </ li> <li> <a href =“ driv. html ”> driver </a> </ li> </ ul>... (omitted)...

  As described above, according to the present embodiment, a company name, a product name, or a unique name of a product can be acquired from an arbitrary Web page.

  13 and 14 are flowcharts illustrating an example of the operation of the rule learning unit 116 according to the present embodiment. The operation will be described below with reference to FIGS.

  (Step S801) The rule learning unit 116 substitutes 0 for the counter m.

  (Step S802) The rule learning unit 116 increments the counter m.

  (Step S803) The rule learning unit 116 branches the process depending on the value of the counter m. When the value of m is 1, the process proceeds to step S804. When the value of m is 2, the process proceeds to step S805. When the value of m is 3, the process proceeds to step S806. When the value of m is 4, the process proceeds to step S807. Proceed to step S810 when the value of m is 5, proceed to step S813 when the value of m is 6, proceed to step S816 when the value of m is 7, and end when the value of m is 8 To do.

  (Step S804) The rule learning unit 116 substitutes “company database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S805) The rule learning unit 116 substitutes “product database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S806) The rule learning unit 116 substitutes “unique name database” for the variable o and “unique name of the product” for the variable d. Then, the process proceeds to step S823.

  (Step S807) The rule learning unit 116 substitutes “company product database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S808) The rule learning unit 116 substitutes “company product database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S809) The rule learning unit 116 substitutes “company product database” for the variable o and “company name / product name” for the variable d. Then, the process proceeds to step S823.

  (Step S810) The rule learning unit 116 substitutes “company unique name database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S811) The rule learning unit 116 substitutes “company unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S812) The rule learning unit 116 substitutes “company unique name database” for the variable o and “company name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S813) The rule learning unit 116 substitutes “product unique name database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S814) The rule learning unit 116 substitutes “product unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S815) The rule learning unit 116 substitutes “product unique name database” for the variable o and “product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S816) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S817) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S818) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S819) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / product name” for the variable d. Then, the process proceeds to step S823.

  (Step S820) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S821) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S822) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S823) The rule learning unit 116 acquires information stored in the variable d from the database stored in the variable o. For example, when “company database” is stored in variable o and “company name” is stored in variable d, all company names are acquired from company database 101. Also, if the information stored in the variable d is “company name / product name”, etc., and two types of information are registered by separating them with “•”, the character string is divided before and after “•”. Get a set of both the back and back information. In the case of “company name / product name”, a set of “company name” and “product name” is acquired. If it can be acquired, the process proceeds to step S824, and if it cannot be acquired, the process proceeds to 829.

  (Step S824) The rule learning unit 116 substitutes 1 for the counter h.

  (Step S825) The rule learning unit 116 determines whether or not there is h-th information in the information regarding d acquired from o acquired in step S823. If there is, the process proceeds to step S826, and if not, the process proceeds to step S831.

  (Step S826) The rule learning unit 116 acquires a Web page from the Web server device. The Web page acquired at this time acquires a Web page including the h-th information regarding d acquired from o acquired in step S823. If the Web page can be acquired, the process proceeds to step S827. If the Web page cannot be acquired, the process proceeds to step S829.

  (Step S827) The rule learning unit 116 replaces the character string of the h-th information related to d acquired from o acquired in Step S823 from the Web page acquired in Step S826 with a character string that understands the content of the information. For example, the company name portion is replaced with “% company name%”.

  (Step S828) The rule learning unit 116 temporarily stores the Web page replaced in Step S827 in a storage medium such as a memory (not shown).

  (Step S829) The rule learning unit 116 increments the counter h.

  (Step S830) The rule learning unit 116 acquires patterns from all the Web pages stored in Step S828. If the pattern can be acquired, the process proceeds to step S831, and if the pattern cannot be acquired, the process proceeds to step S833.

  (Step S831) The rule learning unit 116 selects and acquires patterns extracted five or more times from the patterns acquired in Step S830. If the pattern can be acquired, the process proceeds to step S832, and if the pattern cannot be acquired, the process proceeds to step S833.

  (Step S832) The rule learning unit 116 stores the pattern acquired in step S831 in the rule database corresponding to the variable d. Specifically, if the value of d is “company name”, it is “company rule database”, if it is “product name”, “product rule database”, if it is “unique name”, “unique name rule database”, “ If it is “company name / product name”, “company name / unique name”, “product name / unique name”, it is stored in the “knowledge proliferation rule database”.

  (Step S833) The rule learning unit 116 proceeds to a process next to the process executed before proceeding to step S823.

  Hereinafter, a specific operation of the rule learning unit 116 in the present embodiment will be described. Note that the acquisition result of the rule learning unit 116 shown in this specific example is prepared for convenience of explanation, and does not indicate actual data.

  The rule learning unit 116 selects the company name or the product name from the company database 101, the product database 102, the unique name database 103, the company product database 104, the company unique name database 105, the product unique name database 106, and the company product unique name database 107. Alternatively, a unique name of a product or a combination thereof is acquired as much as possible, and a pattern for extracting information acquired from the database is acquired from a Web page from which the acquired information can be acquired.

  As a specific example, a case will be described in which a rule for acquiring a company name from the company database 101 and storing it in the company rule database 108 is learned.

  First, all company names are acquired from the company database 101. Note that if the company database 101 holds values such as the number of acquisitions and the acquisition ratio of the company name, the company name below an arbitrary threshold value may not be acquired. In this specific example, it is assumed that ten types of company names have been acquired.

  The rule learning unit 116 acquires a web page. For example, in this specific example, assume that 100 types of Web pages are acquired for each acquired company name.

  The rule learning unit 116 replaces the company name used for acquisition with “% company name%” from the acquired Web page. This time, it was “% company name%”, but if the product name is used, it is replaced with “% product name%”, and if the product unique name is used, it is replaced with “% unique name%”. For example, using “sample trading”, “<table> <tbody> <tr> <td> company name </ td> <td> sample trading corporation </ td> </ tr> <tr> <td> capital </ Td> <td> 10,000,000 Yen </ td> </ tr> "in the case where the text" <table> <tbody> <tr> <td> company name </ td> <td> % Company name% corporation </ td> </ tr> <tr> <td> capital </ td> <td> 10,000,000 yen </ td> </ tr>.

  The rule learning unit 116 acquires a pattern from the acquired web page. Two or more Web pages are extracted from the acquired Web pages, and the longest common subsequence including “% company name%” used for acquisition is acquired. Acquisition of the longest common portion is performed by all combinations of selecting two from the acquired Web pages. For example, “<table> <tbody> <tr> <td> Company name </ td> <td>% Company name% Inc. </ Td> </ tr> <tr> <td> Capital </ td> <td > 10,000,000 Yen </ td> </ tr> and "<table> <tr> <td> Company name </ td> <td>% Company name% Inc. </ Td> </ tr> < If there is a Web page with two sentences “tr> <td> Established </ td> <td> January 1, 1950 </ td> </ tr>”, “% company name%” “<Tr> <td> Company name </ td> <td>% Company name% Inc. </ Td> </ tr> <tr> <td>”, which is the longest character string that holds . In this case, since the company name is 10 types and the Web page is 100 types for the company name, the pattern is acquired 49500 times. The pattern acquisition target may be a sentence that does not include an HTML tag, may be a method other than the longest common subsequence, and may acquire a pattern from a combination of two or more Web pages.

  The rule learning unit 116 registers only frequently used patterns from the acquired patterns in the knowledge multiplication database. A pattern that appears at a certain frequency, for example, five times or more is registered in the knowledge proliferation rule database 111.

  As described above, according to the present embodiment, it is possible to acquire a pattern for acquiring a company name, a product name, or a unique name of a product from an arbitrary Web page.

  FIG. 15 is a flowchart showing an example of the operation of the company similar group code unique name database construction unit 23 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 901) The company similar group code unique name database construction unit 23 acquires information on the company name, product name, and product unique name from the company product unique name database 107. If it can be acquired, the process proceeds to step S902. If it cannot be acquired, the process ends.

  (Step S902) The company similar group code unique name database construction unit 23 substitutes 1 for a counter m.

  (Step S903) The company similar group code unique name database construction unit 23 determines whether or not the information acquired in step S901 includes the m-th information. If there is, the process proceeds to step S904, and if not, the process ends.

  (Step S904) The company similar group code unique name database construction unit 23 acquires a similar group code corresponding to the product name from the product similar group code correspondence database 21 using the product name of the m-th information acquired in step S901 as a key. To do.

  (Step S905) The company similar group code unique name database construction unit 23 uses the m-th information company name acquired in step S901, the unique name of the product, and the similar group code acquired in step S904 as one record. Register in the group code unique name database 22.

  (Step S906) The company similar group code unique name database construction unit 23 increments the counter m.

  Hereinafter, a specific operation of the company similar group code unique name database construction unit 23 in the present embodiment will be described. Note that the acquisition result of the company similar group code unique name database construction unit 23 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The company similar group code unique name database construction unit 23 acquires a similar group code from the similar group code correspondence database using the product name of the company product unique name database 107 as a key, and constructs the company similar group code unique name database 22.

  First, the company similar group code unique name database construction unit 23 acquires all the company name, product name, and product unique name from the company product unique name database 107. Note that if the company product unique name database 107 holds values such as the number of acquisitions and the acquisition ratio, the company name below an arbitrary threshold may not be acquired. In this specific example, it is assumed that ten types of combinations of company name, product name, and product unique name have been acquired.

  Next, a similar group code corresponding to each product name is acquired from the product similar group code correspondence database 21. Note that the similar group code acquired at this time is not limited to the complete matching of the product names, but may be a similar group code that conceptually matches the product names such as “TV” and “TV”, for example.

  Then, the set of the acquired company name, similar group code, and product unique name is registered in the company similar group code unique name database 22.

  As described above, according to the present embodiment, the company similar group code unique name that holds the company name, the similar group code, and the unique name of the product using the product name as a key from the company product unique name database 107 and the similar group code correspondence database. A database 22 can be constructed.

  FIG. 16 is a flowchart showing an example of the operation of the trademark infringement detection apparatus 3 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S1001) The receiving unit 32 determines whether or not the investigation target trademark information, specifically, the trademark character string and the product name or the similar group code has been received. If accepted, the process proceeds to step S1002, and if not accepted, the process returns to step S1001.

  (Step S1002) The trademark character string acquisition unit 331 acquires a trademark character string from the investigation target trademark information. If it can be acquired, the process proceeds to step S1003. If it cannot be acquired, the process proceeds to step S1017.

  (Step S1003) The name acquisition unit 332 acquires one or more pieces of name information using the trademark character string acquired in step S1002. For example, a morphological analysis of a trademark character string, or reading of each character constituting the trademark character string is read from a dictionary or the like in which a character reading prepared in advance is stored, and a combination thereof is obtained. Get information.

  (Step S1004) The name acquisition unit 332 substitutes 1 for the counter m.

  (Step S1005) The name acquisition unit 332 determines whether there is m-th name information in the name information acquired in step S1003. If there is, the process proceeds to step S1006, and if not, the process proceeds to step S1011.

  (Step S1006) The name acquisition unit 332 substitutes 1 for the counter n.

  (Step S1007) The name acquisition unit 332 determines whether or not the m-th name information includes the n-th delimiter position. For example, when setting between all characters of the nominal information as the separation position, the nth separation position is, for example, between the nth character and the (n + 1) th character from the beginning of the nominal information. is there. For this reason, the name acquisition unit 332 may determine whether, for example, the mth name information includes the (n + 1) th character from the top. If there is an nth delimiter position, the process proceeds to step S1008, and if not, the process proceeds to step S1010. In the name information, the name information is divided by setting the separation position between reading of each character constituting the trademark character string or reading of the morpheme constituting the trademark character string. May be. For example, when the designation information “ABC” is acquired by combining the readings “A”, “B”, and “C” of each character of the trademark character string “ABC”, the position that can be set as the separation position is “A”. It may be between “Bee” and between “Bee” and “Sea”.

  (Step S1008) The name acquisition unit 332 divides the mth name information at the nth position, and acquires divided name information that is a character string before and after the mth name information. It should be noted that a character string preceded by a predesignated character that cannot be a trademark name is preferably not acquired as divided name information. The character designated in advance is, for example, a long sound or “n”. The acquired divided name information is temporarily stored in a storage medium such as a memory (not shown). If other divided name information is already stored, it is added.

  (Step S1009) The name acquisition unit 332 increments the counter n by 1. Then, the process returns to step S1007.

  (Step S1010) The name acquisition unit 332 increments the counter m by 1. Then, the process returns to step S1005.

  (Step S1011) The name display unit 333 displays the name information acquired in step S1003 and the divided name information acquired in step S1008. Note that when the designation of designation is selected by a specific rule without causing the user to select, the display of this step and the reception of step S1012 may be skipped.

  (Step S1012) The designation designation receiving unit 334 determines whether designation designation information for designating at least one of designation designation information or divided designation designation information has been accepted. If accepted, for example, the designation information and assigned call information specified by the designation designation information are temporarily stored in a storage medium such as a memory, and the process proceeds to step S212. If not received, the process returns to step S1012.

  (Step S1013) The similarity search unit 335 substitutes 1 for the counter k.

  (Step S1014) The similarity search unit 335 determines whether there is k-th name information or divided name information in the combination of the name information and the divided name information specified by the name designation information received in Step S1012. to decide. If there is, the process proceeds to step S1015, and if not, the process proceeds to step S1017.

  (Step S1015) Similarity search is performed using the kth name information or divided name information. Then, the result of the similarity search is stored in a storage medium (not shown). As a result of the similarity search, for example, k-th name information or divided name information used for the similarity search, a name determined to be similar, a company name associated with the name, and a unique name of the product It is information to include.

  (Step S1016) The similarity search unit 335 increments the counter k by 1. Then, the process returns to step S1014.

  (Step S1017) The merchandise kind determination unit 34 selects the database 31 to be used based on the survey target trademark information acquired in step S1001. Specifically, the company similar group code unique name database 22 is selected if the similar group code is included in the investigation target trademark information, and the company product unique name database 107 is selected if the product name is included. . If it can be selected, the process proceeds to step S1018. If it cannot be selected, the process proceeds to step S1019.

  (Step S1018) The product type determination unit 34 determines the product type using the product name or similar group code of the investigation target trademark information acquired from the database 31 of step S1017 in step S1001 as a key. Then, the result of similarity determination is stored in a storage medium (not shown). The result of similarity determination is, for example, the product name or similar group code used for similarity determination, the product name or similar group code determined to match, and the company name and product associated with the designation. Information including the unique name.

  (Step S1019) From the similar search result accumulated in step S1015 and the information accumulated in step S1018, information having a common product unique name is acquired.

  (Step S1020) The output unit 35 outputs the information acquired in step S1019. Then, the process ends.

  Hereinafter, a specific operation of the trademark infringement detection device 3 in the present embodiment will be described. Note that the data shown in this specific example is prepared for convenience of explanation, and is not used in actual operation.

  First, the user wants to determine the possibility of infringement of “wire noodles” for Chinese noodles, and therefore, the trademark character string “wire noodles” and a similar group code “32F03” for Chinese noodles. Is input to the trademark infringement detection device 3 by operating a keyboard or the like. The receiving unit 32 receives the trademark character string “wire noodle”. Then, the trademark character string acquisition unit 331 is caused to acquire a trademark character string, and the similar group code is acquired to the similar product group determination unit 34.

  The name acquisition unit 332 acquires the name information of the trademark character string “wire noodle” received by the trademark character string acquisition unit 331. Here, first, morphological analysis is performed on the trademark character string to obtain reading information for each morpheme. For example, “wire noodle” is divided into two morphemes “wire” and “noodle” by morphological analysis. “Harigane” is read from the morpheme “wire” and “men” is read from the morpheme “noodle”. ”Is acquired. Since only one reading is acquired for each morpheme, “harnessing” obtained by simply combining these readings is acquired as one of the designation information of “wire noodles”.

  Furthermore, the name acquisition unit 332 acquires a reading of each character of the trademark character string “wire noodle” from a dictionary (not shown) in which kanji readings prepared in advance are stored. For example, it is assumed that readings “harness” and “shin” are acquired from “needle”. In addition, it is assumed that “Kin”, “Kogane”, “Kon”, and “Kane” are acquired from “Gold”. It is also assumed that “Men” is acquired from “Noodles”. Then, the name information is acquired by combining the acquired character readings along the sequence of the characters to be acquired. For example, name information such as “Harikinmen”, “Harikomenmen”, “Harikonmen”, “Harikanmen”, “Shinkinmen”, “Shinkoganen”, etc. is acquired. Note that information that matches the name information acquired by morphological analysis is deleted.

  Next, the name acquisition unit 332 acquires the divided name information by dividing the name information acquired above. For example, with respect to the designation information “Hariganemen”, first, by dividing the designation information by setting the division position between “Ha” of the first character and “Li” of the second character, the divided designation of “Ha” is given. Information and division name information “Rigane men” are acquired. Further, by dividing the designation information by setting the division position between the second character “Li” and the third character “Ga”, the divided designation information “Hari” and the divided designation information “Ganemen” And get. Similarly, “Hariga”, “Nemen”, “Harigane”, “Men”, “Harigamime” are acquired. However, “n” is not acquired as divided name information. Similarly, divided name information is acquired for other name information. It should be noted that, in the acquired divided name information, duplicates are deleted except for one.

  Then, the name display unit 333 outputs the acquired name information and divided name information to a monitor or the like.

  FIG. 17 is a diagram showing a display example of the name information and the divided name information by the name display unit 333. In this case, the name information and the divided name information are used without distinction in the later-described similar search, and therefore, the name information and the divided name information are displayed as names without distinction.

  Next, it is assumed that the user checks the check box next to the name information and the divided name information shown in FIG. 17 that he / she wants to use for similarity search and presses the “designation complete” button 31. . For example, if the check boxes next to “Harigamimen”, “Harigami”, and “Harikinmen” are checked as shown in FIG. 17 and the “Designation Complete” button 31 is pressed, the designation designation receiving unit 334 is checked. “Hariganemen” and “Harikinmen”, which are the name information, and “Harigane”, which is the divided name information which is checked, are received.

  Next, the similarity search unit 335 performs a name similarity search using the name information and the divided name information received by the name designation receiving unit 334. Since the similarity search for names is a known technique, a detailed specific example is omitted. For example, as a result of the similarity search, for the name information “Hariganmen” and the divided name information “Harigane”, a registered trademark having a similar name was not detected. However, for the name information “Harikinmen”, a registered trademark having a similar name Suppose that it was detected. Specifically, it is assumed that a unique name of a product whose name is “Hurricington” is detected. The trademark similarity determination result acquisition unit 336 temporarily stores the detection result in a storage medium (not shown).

  Then, the trademark similarity determination result acquisition unit 336 acquires the result of the similarity search.

  Next, information is acquired from the company similar group code unique name database 22 by using the similar group code acquired from the receiving unit 32 for determining the similarity of the product. In the case of “32F03”, information related to the processed grain product is acquired regardless of the “wire surface”, but all information is acquired.

  And among the judgment result of the trademark kind judgment part 33 and the judgment result of the product kind judgment part 34, the information with the common name of goods is acquired, and the company name and the unique name of goods are output. At this time, if there is information for identifying the company in the company database 101, the certainty of the company name, and other held information, it may be output.

  As described above, according to the present embodiment, the name information of the trademark character string is automatically acquired from the survey target trademark character string designated by the user, and the similarity search using this can be performed. It is possible to provide the user with information that can be used to determine the overall infringement status.

  In addition, since the divided name / call information obtained by dividing the name information is acquired and the similarity search using this information can be performed, for example, when the trademark input by the user is a combined trademark, the combined trademark is divided Can be automatically acquired, and an appropriate similarity search using divided name information can be performed. In addition, it is possible to easily input the divided name information and to prompt the user to perform a similar search using the divided name information. In addition, by acquiring the division name information, an appropriate similarity search using only the main part of the trademark becomes possible. Further, it is possible to easily input the divided name information and to prompt the user to perform a similar search using the main part.

(Embodiment 2)
The present embodiment is characterized in that the similarity of the product can be determined using the similar group code even when the investigation target trademark information holds the product name and does not hold the similar group code.

  FIG. 21 is a block diagram of trademark infringement detection apparatus 4 in the present embodiment. The trademark infringement detection apparatus 4 includes a database 401, a reception unit 32, a trademark class rejection determination unit 33, a product class determination unit 402, and an output unit 35. Since the reception unit 32, the trademark similarity determination unit 33, and the output unit 35 are the same as those in the first embodiment, description thereof is omitted.

  The database 401 can store one or more sets of company names, product names, and product unique names, which are information acquired from one or more Web server devices (not shown). The database 401 may be any information obtained from one or more Web server devices, may be constructed using the database construction device 1, and is searched by using a product name of random or surveyed trademark information as a keyword. Information may be acquired from a Web page collected using a specific rule. The specific rule is, for example, a rule stored in each rule database of the database construction device 1.

  The commodity similarity determination unit 402 performs commodity similarity determination using one or more commodity names included in the database 401 and the code identification information included in the investigation target trademark information, and the investigation target stored in the database 401 The company name, product name, and product unique name similar to the trademark information are acquired. The product category determination unit 402 includes a product code management unit 4021, a code identification information acquisition unit 4022, and a product category determination unit 4023. The code specifying information in the present embodiment may be a product name, a number for referring to a similar group code managed in a storage unit (not shown), or the like, and a similar group code table May be the product or service name (for example, “Chemical” in the case of 01A01), the product / service name in the similar group code table, or the similar group code itself.

  The product code management unit 4021 manages the correspondence between the product name and the code specifying information. The product name to be managed and the code specifying information may be acquired from the description of the similar group code using a specific rule, or may be acquired from a Web page acquired from a Web server device (not shown) using a specific rule. Often, a person may input manually. The specific rule is, for example, a rule for acquiring “04A01” as the code identification information and “washing powder” and “shampoo” as the product name from a sentence such as “04A01 is for example washing powder or shampoo”. From the HTML structure such as “<table> <tr> <td> dye </ td> <td> indigo / akane </ td> </ tr> </ table>” as code specifying information The rule may be to acquire “dye” and “indigo” and “Akane” as trade names. The product code management unit 4021 may be configured to acquire a similar group code by recursively searching for the correspondence. Specifically, when there are two sets of information, code specific information “01A01” and product name “Chemical”, code specific information “Chemical” and product name “Ammonia water”, the product name “Ammonia water” You may manage so that code specific information "01A01" may be acquired.

  The code specifying information acquisition unit 4022 acquires code specifying information corresponding to one or more product names included in the database 401, using the code specifying information corresponding to the product managed by the product code management unit 4021. The code specifying information to be acquired may be arbitrary code specifying information corresponding to the product name of the database 401, may be a similar group code acquired as a result of recursively acquiring the correspondence, It may be all the code specifying information acquired in the process of acquiring the correspondence recursively until it arrives. The code specifying information acquisition unit 4022 can be usually realized by an MPU, a memory, or the like. The processing procedure of the code specifying information acquisition unit 4022 is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The product type determination unit 4023 makes a product type determination using the code specification information acquired by the code specification information acquisition unit 4022 and the code specification information included in the investigation target trademark information. In determining whether a product is similar, the code specifying information included in the trademark information to be investigated and the code specifying information acquired by the code specifying information acquiring unit 4022 are determined to be the same or included. In this case, the company name, the product name, and the unique name of the product corresponding to the information are acquired. The merchandise kind determination unit 4023 can usually be realized by an MPU, a memory, or the like. The processing procedure of the merchandise kind determination unit 4023 is usually realized by software, and the software is recorded in a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The operation of the trademark infringement detection apparatus 4 according to the present embodiment is the same as the operation of the trademark infringement detection apparatus 3 except for the processing in step S1018 in FIG.

  Hereinafter, a specific operation of the merchandise inequality determination unit 402 in the trademark infringement detection device 4 according to the present embodiment will be described. It should be noted that the output result of the commodity similarity determination unit 402 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data. Here, as an example, the code specifying information of the investigation target trademark information received by the receiving unit 32 is “beer”, and the database 401 has a company name “A company”, a product name “Lugger beer”, and a unique product name. It is assumed that at least one set of information whose name is “B beer” is stored, and the product code management means 4021 has the code specifying information “28A02” and the product name “beer”, the code specifying information “beer” and the product name “ It is assumed that at least two sets of information “Lager beer” are stored.

  The code specifying information acquisition unit 4022 acquires information on the product name from the database 401. For example, information such as “Lugger beer” is acquired. Then, using the information managed by the product code management means 4021, the code specifying information is acquired from each product name. Specifically, all code specifying information from “Lager Beer” to “Beer” and “Beer” to “28A02” is acquired and stored in a storage unit (not shown) in association with the product name. To do.

  The product type determination unit 4023 acquires the information holding the code specifying information identical to the “beer” included in the survey target trademark information received by the receiving unit 32 by the code specifying information acquiring unit 4022, and stores it in a storage unit (not shown). Searching from the stored information, the company name and the unique name of the product stored in the database 401 are acquired. Specifically, “Lager Beer” associated with “Beer” and “28A02” is acquired, the company name is “Company A”, the product name is “Lager Beer”, and the product unique name is “B Beer”. Is obtained.

  As described above, according to the present embodiment, a trademark taking account of a similar group code without specifying a similar group code for a set of a company name, a product name, and a product unique name acquired from a Web server device. Thus, it is possible to provide the user with information that can be used to determine the trademark infringement status.

  In the present embodiment, the code specifying information is acquired by the code specifying information acquiring unit 4022, but the code specifying information may be acquired in advance using a processing unit (not shown) and stored in the database 401. good.

  In the first and second embodiments, the company name acquisition unit 112, the product name acquisition unit 113, the unique name acquisition unit 114, the knowledge acquisition unit 115, the rule learning unit 116, and the company similar group code unique name database construction Unit 23, product category determination unit 34, trademark character string acquisition unit 331, name acquisition unit 332, similarity search unit 335, trademark class determination result acquisition unit 336, code identification information acquisition unit 4022, product category determination unit 4023 Usually, it can be realized from an MPU, a memory, or the like. The processing procedure of the product name acquisition unit 113 and the like is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

(Embodiment 3)
In the present embodiment, a database construction apparatus 5 that acquires two or more attribute value sets from a Web page and generates a new attribute value set from the two or more attribute value sets will be described.

  In the present embodiment, the storage unit or the like in which predetermined information can be stored does not matter how the information is stored in the storage unit. For example, information may be stored in a storage unit or the like via a recording medium. Information transmitted via a communication line or the like may be stored in a storage unit or the like. Furthermore, information input via the input device may be stored in a storage unit or the like. In addition, the format, content, and the like of each information described in this embodiment are merely examples, and the format, content, and the like are not limited as long as the meaning of each information can be indicated.

  FIG. 21 is a block diagram of the database construction device 5 in the present embodiment. The database construction device 5 includes a table storage unit 51, a rule storage unit 52, a score storage unit 53, a dictionary storage unit 54, an acquisition unit 55, a combining unit 56, and a rule acquisition unit 57.

  The table storage unit 51 stores a first table and two or more second tables. The first table has one or more attribute value sets having m attribute values (m ≧ 3) having different attributes. Further, the second table has one or more attribute value sets having n (n ≦ m−1) attribute values having different attributes. That is, the attribute value set is a record in one data model. The first table and the second table are a set of one or more records in one data model. For example, when the data model is a relational data model, the first table and the second table are tables. The contents of attributes and attribute values are not limited.

  Also, the first table can usually be the second table. For example, when m is 3, a table having one or more attribute value sets having three attribute values having different attributes is the first table, and two or one attribute values having different attributes are A table having one or more attribute value sets is a second table. When m is 4, a table having one or more attribute value sets having four attribute values having different attributes is the first table, and three, two, or one attribute having different attributes. A table having one or more attribute value sets having values is the second table. That is, whether the table stored in the table storage unit 51 is the first table or the second table is relatively determined by the relationship with other tables.

  The “attribute value of the attribute” means “attribute value corresponding to the attribute”. The “attribute value attribute” means “attribute corresponding to the attribute value”. Further, “attribute value set of attribute” means “attribute value set corresponding to attribute”, and means an attribute value set having one or more attribute values of the attribute. The “attribute value set attribute” means “attribute corresponding to the attribute value set”, and means one or more attribute value attributes of the attribute value set.

  The second table may have a unique identifier indicating whether each of the n attributes is unique. The unique identifier usually has each attribute, but the unique identifier may indicate which attribute is unique. The unique identifier is, for example, a unique key in so-called RDB, but may be a primary key. The unique identifier may indicate whether each of the two or more second tables is unique.

  Further, the first table and the second table may have scores corresponding to the attribute value sets. The score usually indicates the reliability of the attribute value set. Further, the score may correspond to each attribute value included in the attribute value set.

  The “reliability of attribute value set” indicates the degree of trust for the attribute value set. “The reliability of the attribute value set” is, for example, the reliability of the Web page that acquired the attribute value set, the reliability of the attribute value set itself, or the like. “Web page reliability” is, for example, high when the Web page is a so-called official site Web page, and low when the Web page is a web page other than the so-called official site. The “reliability of the attribute value set itself” is, for example, the reliability of the Web page from which the attribute value set is acquired, the attribute values of the attribute value set, and the attribute that is the condition of the attribute value The degree of coincidence with the value condition.

  The attribute value condition is, for example, an attribute value example, an attribute value pattern, or the like. In other words, the “degree of matching with the attribute value condition” means the degree of matching between each attribute value of the attribute value set and an example of the attribute value of the same attribute as the attribute value, This is the degree of coincidence between each attribute value and the attribute value pattern of the same attribute as the attribute value. An example of the attribute value is the attribute value itself. The attribute value pattern includes, for example, that the attribute value includes a predetermined character, that the attribute value is a kanji string, that the attribute value is a katakana string, and that the attribute value is a hiragana string. , Indicating that the attribute value is a numerical value.

  The “attribute value set reliability” may be called “attribute value set accuracy” or “attribute value set accuracy”. The score is a score corresponding to the attribute value set or each attribute value of the attribute value set, and can determine whether the attribute value set or the attribute value is reliable or correct. If so, the content is not questioned.

  The rule storage unit 52 stores one or more rules. The rule is a rule for acquiring one or more attribute value sets having the m or n attribute values from one or more Web pages. The rule is normally used by the acquisition unit 55 described later, but may be any rule as long as it can acquire one or more attribute value sets from one or more Web pages.

  The rule is a rule for obtaining one or more attribute value sets from a text document (for example, a web page displayed by a so-called web browser) obtained by interpreting HTML (Hyper Text Markup Language) included in the web page, For example, a rule for acquiring one or more attribute value sets from a Web page which is a so-called HTML document.

  A rule is usually a regular expression. For example, a rule is a so-called character string pattern, tag pattern, URI (Uniform Resource Identifier) pattern, or a combination of two or more of these three patterns. There may be. Further, the rules may be other notations and expressions. A rule usually has a variable indicating that an attribute value is acquired. The variable corresponds to an attribute.

  For example, if the Web page is a text document, the rules are (1) “$ company name $ is. *? $ Product name $, $ proprietary name $” or (2) “% unique name %:% Product Name% ”, (3)“ Company Overview <New Line> Company Name ”& Company Name & <New Line> Established | & Date of Establishment &&. (1) is a rule for acquiring a company name, a product name, and a unique name. In (1), for example, “$ company name $” is a variable indicating that a character string at the location is acquired as the company name. “. *?” Is a regular expression. Moreover, (2) is a rule for acquiring a unique name and a product name. In (2), for example, “% unique name%” is a variable indicating that a character string at the location is acquired as a unique name. (3) is a rule for acquiring the company name and the date of establishment. In (3), for example, “& date of establishment &” is a variable indicating that a character string at the location is acquired as the date of establishment. “<Line feed>” indicates a line feed.

  Also, for example, if the Web page is an HTML document, the rule is (4) “<p> $ company name $ is. *? $ Product name $, $ unique name $ </ p>” (5) “<td>% unique name%: </ td> <td>% product name% </ td>”, (6) “<div. *?> Company profile </ div> <br> <Div. *?> Company name </ div> <div. *?> & Company name & </ div> <br> <div. *?> Establishment </ div> <div. *?> & Date of establishment & </ Div> ". In addition, since the meaning of said (4)-(6) is the same as said (1)-(3), description is abbreviate | omitted. Further, the meanings of variables and symbols in the rules (4) to (6) are the same as those in the above (1) to (3), and thus the description thereof is omitted.

  The rule storage unit 52 may store patterns corresponding to the variables. The pattern is, for example, that a character string indicated to be acquired by the variable includes a predetermined character, that the character string is a kanji string, that the character string is a katakana string, This indicates that the character string is a hiragana string, the character string is a numerical value, or the like.

  In addition, the Web page is usually included in one or more Web sites, and is included in a server device that operates the Web site. The web page may be stored in any storage device or storage area provided in the database construction device 5. Web pages include documents such as the above-mentioned text documents and HTML documents, sentences, paragraphs, characters, character strings, texts, and the like, and are widely understood.

  The web page may have a web page identifier for identifying the web page. The web page identifier includes the URL of the web page, the file name of the web page, the title of the web page (element value of the HTML title element), meta information (keyword attribute value or description attribute value of the HTML meta element) Etc. The URL includes a so-called domain name and host name and is widely understood.

  Further, the rule is usually stored in the rule storage unit 52 in association with the first table or the second table having an attribute value set that can be acquired by the rule. This means that the correspondence between the rule and one or more attributes corresponding to the attribute value set that can be acquired by the rule or a table that can have the attribute value set is stored so as to be understood.

  The score storage unit 53 stores one or more scores associated with Web page identifiers. The score is a score of the web page, and usually indicates the reliability of the web page.

  The dictionary storage unit 54 stores one or more dictionaries that are a set of one or more attribute values for each attribute. A dictionary may be thought of as a collection of example attribute values. One dictionary is usually a set of one or more attribute values of one attribute. The dictionary storage unit 54 may store one or more attribute value patterns. In this case, the attribute value pattern is usually associated with the attribute value attribute.

  The acquisition unit 55 acquires one or more attribute value sets from one or more Web pages according to one or more rules stored in the rule storage unit 52. At this time, the acquisition unit 55 normally acquires a character string that matches the one or more rules from the one or more Web pages. And the acquisition part 55 acquires the character string in the location of the 1 or more variable which the used rule has as an attribute value from the acquired character string. Then, the acquisition unit 55 accumulates the acquired one or more attribute value sets in the table storage unit 51. The accumulation is to add to the first table or the second table. The additional writing includes the meaning of addition or insertion.

  For example, when one or more attribute value sets having the m attribute values are acquired according to the rule, the acquisition unit 55 adds the acquired attribute value set to a table having attributes corresponding to the attribute value set. . This table is usually the first table. Further, for example, when one or more attribute value sets having the n attribute values are acquired according to the rule, the acquisition unit 55 stores the acquired attribute value sets in a table having attributes corresponding to the attribute value sets. Append. This table is usually the second table.

  Further, the acquisition unit 55 is an attribute value of an attribute corresponding to the variable of one or more rules stored in the rule storage unit 52, and one or more attributes stored in the table storage unit 51. A value may be substituted to generate a new rule, and one or more attribute value sets may be acquired from one or more Web pages according to the rule.

  In addition, the acquisition unit 55 substitutes the pattern stored in the rule storage unit 52 for the variable of one or more rules stored in the rule storage unit 52 and substitutes the pattern stored in the rule storage unit 52 for a new rule. One or more attribute value sets may be acquired from one or more Web pages according to the rules generated.

  The acquisition unit 55 normally performs so-called crawling and acquires one or more Web pages from a so-called Web. For example, when one or more Web pages are stored in an arbitrary storage area, the acquisition unit 55 acquires one or more Web pages from the storage area.

  For example, a character string “PATENT ANALYZER has released software“ PAT-Analyzer ”for analyzing patent specifications” exists in one or more Web pages, and the rule is “$ company name $ is. *? If $ product name $ “$ unique name $” has been released ”, the acquisition unit 55 first acquires the character string that matches the rule. Then, the acquisition unit 55 uses the variable “$ company name $” for acquiring the attribute value from the character string, the character strings “patent analysis corporation”, “software” in the place of “$ product name $”. , “PAT-Analyzer” are acquired as the company name, product name, and unique name, respectively. And the acquisition part 55 adds the acquired attribute value set to the table | surface which has three attributes of a company name, a product name, and a specific name.

  For example, when the company name “Patent Analysis Co., Ltd.” is already stored in the table storage unit 51, the acquisition unit 55 substitutes the company name into the above rule, and a new rule “Patent Analysis Co., Ltd. . *? $ Product name $ "$ Unique name $" was released ", and using this rule, the character strings" Software "and" PAT-Analyzer "were acquired as the product name and unique name, respectively, as described above May be. And the acquisition part 55 adds the acquired attribute value set to the table | surface which has three attributes of a product name and a unique name.

  For example, in the above case, when the pattern corresponding to the variable “$ company name $” is “<Kanji string> corporation”, the acquisition unit 55 sets “<Kanji string> corporation” in the rule. Substituting a new rule “<Kanji> *? $ Product name $ "$ Unique name $" released "is generated. Then, the acquisition unit 55 acquires the character string that matches the rule. Then, the acquisition unit 55 uses the variable “$ company name $” for acquiring the attribute value from the character string, the character strings “patent analysis corporation”, “software” in the place of “$ product name $”. , “PAT-Analyzer” are acquired as the company name, product name, and unique name, respectively.

  In addition, for example, the acquisition unit 55 may acquire a Web page identifier included in the Web page that acquired the attribute value set when acquiring the attribute value set. In this case, the acquisition unit 55 associates the attribute value set with the Web page identifier and adds the attribute value set to the first table or the second table.

  For example, the acquisition unit 55 may calculate a score corresponding to the attribute value set when acquiring the attribute value set. In this case, the acquisition unit 55 associates the attribute value set with the score and adds the attribute value set to the first table or the second table. Moreover, the acquisition part 55 may calculate the score corresponding to each attribute value which the acquired attribute value set has, for example. In this case, the acquisition unit 55 associates each attribute value with the score and adds the attribute value to the first table or the second table. In addition, the acquisition unit 55 may calculate a score corresponding to the attribute value set using a score corresponding to the attribute value. The calculation is, for example, calculating an average score corresponding to each attribute value. In this case, the acquisition unit 55 associates the attribute value set with the score and adds the attribute value set to the first table or the second table.

  For example, the acquisition unit 55 may acquire a score corresponding to the Web page identifier included in the Web page from which the attribute value set has been acquired from the score storage unit 53. For example, when the URL of the Web page from which the attribute value set is acquired is “http://www.webpage.com”, and the URL is associated with the score “95” in the score storage unit 53, the acquisition unit 55 The score “95” is acquired. Then, the acquisition unit 55 associates the attribute value set with the score and adds the attribute value set to the first table or the second table.

  Further, for example, the acquisition unit 55 may calculate the degree of coincidence between the acquired attribute value set and the attribute value condition as a score. In this case, the acquisition unit 55 normally calculates the degree of coincidence between each attribute value included in the attribute value set and the attribute value condition. In this case, the acquisition unit 55 normally calculates the degree of coincidence with the attribute value condition of the same attribute as the attribute of each attribute value.

  For example, assume that the attribute value of the attribute value set is the company name “Patent Analysis Co., Ltd.”. In this case, when the attribute value condition is the company name example “Patent Analysis Co., Ltd.”, the company name and the company name example match, so the acquisition unit 55 acquires the matching degree “100”. Further, in this case, when the attribute value condition is an example of the company name “Patent Analysis Co., Ltd.”, the acquisition unit 55 sets the number of characters “7” that matches the example of the company name among the characters of the company name. Then, the number of characters “8” of the company name is acquired, “7 ÷ 8” is calculated, and the matching degree “0.875” is acquired. Further, in this case, when the attribute value condition is the company name pattern “Kanji character string”, since the company name is a Kanji character string, the acquisition unit 55 acquires the matching degree “100”. In this case, when the attribute value condition is the company name pattern “katakana string”, the acquisition unit 55 acquires the degree of match “0” because the company name does not include any katakana characters.

  When the attribute value condition is an example of an attribute value, the example of the attribute value is an attribute value normally stored in the dictionary storage unit 54. When the attribute value condition is an attribute value pattern, the attribute value pattern is normally held by the acquisition unit 55, but may be stored in an arbitrary storage area or a predetermined storage area. Good.

  For example, the acquisition unit 55 may search the acquired attribute value set with a so-called search engine and acquire the number of Web pages having the attribute value set as a score. For example, the acquired attribute value set is a company name “Patent Analysis Co., Ltd.” and a product name “PAT-Analyzer”, and a search engine searches for a Web page having both of the two attribute values. , ”Indicates that the number“ 1,000 ”is acquired as a score.

  In addition, the acquisition unit 55 may acquire, for example, the cumulative acquisition count of the acquired attribute value set as a score. Since the method of acquiring the cumulative number of times of acquisition is self-evident, detailed description is omitted.

  The combining unit 56 combines two or more attribute value sets that the second table has, and generates a new attribute value set that the first table can have. Then, the combining unit 56 adds the generated attribute value set to the first table. Here, “the first table can have” means that it can be a set of attribute values that the first table has, and means that it has the m attribute values. Specifically, it means that the number of attributes is larger than the number of attributes corresponding to two or more attribute value sets used when generating the attribute value set.

  For example, it is assumed that the combining unit 56 generates the attribute value set having the m attribute values by combining the two attribute value sets having the n attribute values. In this case, the number of attributes in the attribute value set (m) generated by the combination is larger than the number of attributes in the attribute value set to be combined (n). Therefore, the combining unit 56 adds the generated attribute value set to the first table. For example, it is assumed that the combining unit 56 combines two attribute value sets having three attribute values having different attributes, and generates an attribute value set having three attribute values having different attributes. In this case, the number of attributes in the attribute value set to be combined (three) is the same as the number of attributes in the attribute value set generated by combining (three). Therefore, the combining unit 56 does not add the generated attribute value set to the first table. In this case, the combining unit 56 usually compares the number of attributes in the attribute value set to be combined with the number of attributes in the attribute value set generated by combining, and the latter number rather than the former number. If there are more, a new attribute value is generated.

In addition, the combining unit 56 generates a new attribute value set that the first table may have, for example, by the following method.
(1) Two or more attribute value sets that each of two or more different second tables have, have the same attribute value of the same attribute, and that the attribute is unique by the unique identifier Combine attribute value sets.
(2) An attribute value set included in each of three or more different second tables, and attribute value sets having two or more identical attribute values of the same attribute are combined.

  The term “join” is synonymous with the so-called RDB table join (JOIN). In other words, combining means that a single attribute value (key) is duplicated in the attribute value set to be combined and is associated with the remaining attribute values to generate a new attribute value set.

(1) is to combine two or more attribute value sets that meet the following conditions.
(A) It has the same attribute as the attribute of another attribute value set.
(B) The attribute value of the attribute of (a), which has the same attribute value as the attribute value of another attribute value set.
(C) The attribute of (a) is indicated by a unique identifier.

  Note that (1) is, for example, a combination method in which all of the attribute value sets to be combined have a unique identifier, and the above (a) to (c) This is a condition for performing the coupling (joining condition). In (1), two or more attribute value sets are combined using an attribute indicated by the unique identifier as a key.

(2) is to combine three or more attribute value sets that satisfy the following conditions.
(A) It has two or more attributes that are the same as the attributes of the other two or more attribute value sets.
(B) The attribute value of the attribute of (a), which has the same attribute value as the attribute value of another attribute value set.

  Note that (2) is a combining method when, for example, at least one of the attribute value sets to be combined does not have a unique identifier, and the above (a) and (b) Is a condition (coupling condition) for performing the coupling. Further, when performing the combination of (2), the combining unit 56 is a set of three or more attribute values, and the correspondence between two attributes of the three or more attribute value sets is the three or more attribute value sets. It is preferable to combine attribute value sets that are apparent from

  For example, it is assumed that there are three attribute value sets “AB”, “BC”, and “CA” having any of the attribute values “A”, “B”, and “C”. Each of these three attribute value sets has two or more attribute values that are the same as the attribute values of the other two attribute value sets. Therefore, the combining unit 56 combines these three attribute value sets to generate a new attribute value set “ABC”.

  Also, for example, three attribute value sets “ABC”, “CD”, “DA” having any of the attribute values “A”, “B”, “C”, “D” Suppose there is. Each of these three attribute value sets has two or more attribute values that are the same as the attribute values of the other two attribute value sets. Accordingly, the combining unit 56 combines these three attribute value sets to generate a new attribute value set “ABCD”. In this case, the correspondence between “B” and “D” is not clear from these three attribute value sets. Therefore, it is preferable that the combining unit 56 does not combine these three attribute value sets.

  For example, there are three attribute value sets “AB”, “BC”, and “CD” having any of the attribute values “A”, “B”, “C”, and “D”. And Of these three attribute value sets, “CD” does not have two or more attribute values that are the same as the attribute values of the other two attribute value sets. Accordingly, the combining unit 56 does not combine these three attribute value sets. In other words, this is for preventing the correspondence between “A” and “D” from being established in a state where the correspondence between “A” and “D” is not clear.

  Further, for example, three attribute value sets “ABC” and “ABD” having any of the attribute values “A”, “B”, “C”, “D”, “E” And “C-D-E”. Each of these three attribute value sets has two or more attribute values that are the same as the attribute values of the other two attribute value sets. Accordingly, the combining unit 56 combines these three attribute value sets to generate a new attribute value set “ABCDE”. In this case, the correspondence between “E” and “A” and “E” and “B” is not clear from these three attribute value sets. Therefore, it is preferable that the combining unit 56 does not combine these three attribute value sets. However, the combining unit 56 combines the attribute value sets that are part of these three attribute value sets and have a clear correspondence, and generates a new attribute value set “ABCD”. Also good.

  In addition, although the above demonstrated only the attribute value for convenience, normally, unless the attribute of the said attribute value is also the same, combining will not be performed.

  Further, for example, the combining unit 56 may generate a new attribute value set using only an attribute value set whose score corresponding to the attribute value set is high enough to satisfy a predetermined condition. The condition is normally held by the coupling unit 56, but may be stored in an arbitrary storage area.

  The combining unit 56 may combine two or more attribute value sets included in the second table to generate a new attribute value set that the first table may have. It doesn't matter. Further, the attribute value set of the second table that is the source of the generation may be anything as long as it is stored in the table storage unit 51. For example, the combining unit 56 may combine two or more attribute value sets acquired by the acquiring unit 55, may combine two or more attribute value sets generated by the combining unit 56, or may acquire the acquiring unit 55. Two or more attribute value sets from among the attribute value sets acquired by or the attribute value set generated by the combining unit 56 may be combined.

  The rule acquisition unit 57 first acquires a character string including one or more attribute value sets stored in the table storage unit 51 from one or more Web pages. And the rule acquisition part 57 acquires character strings other than the said attribute value set among the said character strings as said rule. Then, the rule acquisition unit 57 accumulates the acquired rules in the rule storage unit 52.

  The rule acquisition unit 57 performs, for example, so-called crawling and acquires one or more Web pages. Then, the rule acquisition unit 57 normally acquires a Web page having an attribute value set from the Web page. Further, the rule acquisition unit 57 may acquire one or more Web pages using, for example, a so-called search engine using an attribute value set as a search key.

  Also, the rule acquisition unit 57 acquires a character string that can be said to be a single unit, such as a sentence, row, or table in which an attribute value set exists, from one or more Web pages acquired as described above.

  For example, the rule acquisition unit 57 replaces the attribute value set existing in the character string acquired as described above with a predetermined variable. Then, the rule acquisition unit 57 acquires the replaced character string as a rule. For example, when a company name exists in the character string, the rule acquisition unit 57 replaces the company name existing in the character string with a variable (for example, “% company name%”) corresponding to the company name. For example, when the product name and the unique name exist in the character string, the rule acquisition unit 57 converts the product name and the unique name that exist in the character string into variables (for example, the product name and the unique name, for example, “$ Product name”, “$ unique name $”).

  Moreover, the rule acquisition part 57 may replace the character string whose object is self-evident with a predetermined variable, for example. The “character string whose target is self-evident” is, for example, a date, time, day of the week, name of person, and the like. Moreover, the rule acquisition part 57 may replace character strings other than what is called a verb, a particle, a noun, etc. with the symbol which shows arbitrary 0 or more character strings, for example.

  In addition, the rule acquisition unit 57 may generate a new rule from the acquired rule and the rule already stored in the rule storage unit 52, for example. In this case, the rule acquisition unit 57 normally acquires from the rule storage unit 52 a rule that is so high that the degree of similarity with the acquired rule satisfies a predetermined condition, and sets a common part of the two rules as a new one. Get as a rule. At this time, any part other than the common part may be replaced with an arbitrary symbol indicating zero or more character strings, or may be connected by so-called logical sum. The rule acquisition unit 57 may create a new rule as described above, for example, from two or more acquired rules or from two or more rules already stored in the rule storage unit 52.

  Also, the rule acquisition unit 57 accumulates the rules acquired as described above in the rule storage unit 52. At this time, the rule acquisition unit 57 normally accumulates the acquired rule in the rule storage unit 52 in association with the first table or the second table having the attribute value set used when the rule is acquired. To do. Note that the rule acquisition unit 57 may accumulate the acquired rules in the rule storage unit 52 so that the correspondence between the acquired rules and the attribute value set (attribute set) that can be acquired by the rule is understood. Any method or procedure is acceptable.

  The table storage unit 51, the rule storage unit 52, the score storage unit 53, and the dictionary storage unit 54 are preferably non-volatile recording media, but can also be realized by volatile recording media.

  In addition, the acquisition unit 55, the combining unit 56, and the rule acquisition unit 57 can be usually realized by an MPU, a memory, or the like. The processing procedure of the acquisition unit 55 and the like is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  Next, the operation of the database construction device 5 will be described. Note that the i-th information in the predetermined information is described as “information [i]”. FIG. 22 is a flowchart showing the overall operation of the database construction device 5.

  (Step S2201) The acquisition unit 55 determines whether to acquire an attribute value set. Specifically, for example, the database construction device 5 has a reception unit (not shown), and determines whether or not the reception unit has received an instruction to acquire an attribute value set. In addition, the timing (cycle) for acquiring the attribute value set is determined in advance, and it is determined whether or not the timing is reached. When acquiring, it progresses to step S2202, and when that is not right, it progresses to step S2203.

  (Step S2202) The acquisition unit 55 acquires an attribute value set. Details of this processing will be described with reference to the flowchart of FIG. Then, the process returns to step S2201.

  (Step S2203) The combining unit 56 determines whether to generate an attribute value set. Specifically, for example, the database construction device 5 has a reception unit (not shown), and determines whether or not the reception unit has received an instruction to generate an attribute value set. The timing (cycle) for generating the attribute value set is determined in advance, and it is determined whether or not the timing is reached. If so, the process proceeds to step S2204; otherwise, the process proceeds to step S2205.

  (Step S2204) The combining unit 56 generates an attribute value set. Details of this processing will be described with reference to the flowchart of FIG. Then, the process returns to step S2201.

  (Step S2205) The rule acquisition unit 57 determines whether or not to acquire a rule. Specifically, for example, the database construction device 5 has a reception unit (not shown), and it is determined whether or not the reception unit has received an instruction to acquire a rule. The timing (cycle) for acquiring the rule is determined in advance, and it is determined whether or not the timing is reached. When acquiring, it progresses to step S2206, and when that is not right, it returns to step S2201.

  (Step S2206) The rule acquisition unit 57 acquires a rule. Details of this processing will be described with reference to the flowchart of FIG. Then, the process returns to step S2201.

  In the flowchart of FIG. 22, the process may be terminated by powering off or a process termination interrupt.

  In the flowchart of FIG. 22, the order of the determination in step S2201, the determination in step S2203, and the determination in step S2205 does not matter.

  FIG. 23 is a flowchart showing the attribute value set acquisition processing in step S2202 of the flowchart of FIG.

  (Step S2301) The acquisition unit 55 acquires all the rules stored in the rule storage unit 52. Here, it is assumed that M rules have been acquired.

  (Step S2302) The acquisition unit 55 performs crawling and acquires one or more Web pages from the Web. Here, it is assumed that N Web pages have been acquired.

  (Step S2303) The acquisition unit 55 sets 1 to the counter i.

  (Step S2304) The acquisition unit 55 sets 1 to the counter j.

  (Step S2305) The acquisition unit 55 sets 0 to the variable count.

  (Step S2306) The acquisition unit 55 determines whether a character string that matches the rule [i] exists in the Web page [j]. When it exists, it progresses to step S2307, and when that is not right, it progresses to step S2311.

  (Step S2307) The acquisition unit 55 acquires a character string that matches the rule [i] from the Web page [j].

  (Step S2308) The acquisition unit 55 acquires an attribute value set from the acquired character string.

  (Step S2309) The acquisition unit 55 increments count by 1.

  (Step S2310) The acquisition unit 55 sets the acquired attribute value set in the variable attrs [count].

  (Step S2311) The acquisition unit 55 determines whether j is N or not. When it is N, it progresses to step 2313, and when that is not right, it progresses to step S2312.

  (Step S2312) The acquisition unit 55 increments j by 1, and returns to step S2306.

  (Step S2313) The acquisition unit 55 determines whether count is 1 or more. If it is 1 or more, the process proceeds to step S2314; otherwise, the process proceeds to S2315.

  (Step S2314) The acquisition unit 55 adds attrs to the second table.

  (Step S2315) The acquisition unit 55 determines whether i is M. If it is M, the process returns to the upper process, and if not, the process proceeds to step S2316.

  (Step S2316) The acquisition unit 55 increments i by 1, and returns to Step S2304.

  In the flowchart of FIG. 23, one or more attribute value sets are acquired from N Web pages for each rule. Thereby, for each attribute of the attribute value set acquired using the rule, the acquired attribute value set can be added to the second table having the attribute.

  In the flowchart of FIG. 23, for all M rules, one or more attribute value sets are acquired from N Web pages, and all the acquired attribute value sets are collectively displayed in the second table. You may add.

  FIG. 24 is a flowchart showing the attribute value set generation processing in step S2204 of the flowchart of FIG.

  (Step S <b> 2401) The combining unit 56 acquires all the second tables stored in the table storage unit 51. Here, it is assumed that M attribute value sets have been acquired.

  (Step S2402) The combining unit 56 sets 1 to the counter i.

  (Step S2403) The combining unit 56 determines whether any of the attributes corresponding to the attribute value set [i] has a unique identifier. If so, the process proceeds to step S2404; otherwise, the process proceeds to step S2413.

  (Step S2404) The combining unit 56 acquires an attribute having a unique identifier and an attribute value corresponding to the attribute from the second table having the attribute value set [i] and the attribute value set [i].

  (Step S2405) The combining unit 56 sets 1 to the counter j.

  (Step S2406) The combining unit 56 determines whether i and j are different. If they are different, the process proceeds to step S2407; otherwise, the process proceeds to step S2409.

  (Step S2407) The combining unit 56 determines whether or not the attribute value set [j] has an attribute value corresponding to the attribute acquired in step S2404 and the same as the attribute value acquired in step S2404. If so, the process proceeds to step S2408; otherwise, the process proceeds to step S2409.

  (Step S2408) The combining unit 56 combines the attribute value set [i] and the attribute value set [j].

  (Step S2409) The combining unit 56 determines whether j is M. When it is M, it progresses to step S2411, and when that is not right, it progresses to step S2410.

  (Step S2410) The combining unit 56 increments j by 1, and returns to step S2406.

  (Step S2411) The combining unit 56 determines whether i is M. If it is M, the process returns to the upper process, and if not, the process proceeds to step S2412.

  (Step S2412) The combining unit 56 increments i by 1, and returns to step S2403.

  (Step S2413) The combining unit 56 sets 1 to j.

  (Step S2414) The coupling unit 56 sets 1 to the counter k.

  (Step S2415) The combining unit 56 determines whether i, j, and k are different. If they are different, the process proceeds to step S2416; otherwise, the process proceeds to step S2418.

  (Step S2416) The combining unit 56 determines whether or not the attribute value set [i], the attribute value set [j], and the attribute value set [k] satisfy the connection condition. If so, the process proceeds to step S2417; otherwise, the process proceeds to step S2418.

  (Step S2417) The combining unit 56 combines the attribute value set [i], the attribute value set [j], and the attribute value set [k].

  (Step S2418) The combining unit 56 determines whether or not k is M. When it is M, it progresses to step S2420, and when that is not right, it progresses to S2419.

  (Step S2419) The combining unit 56 increments k by 1, and returns to step S2415.

  (Step S2420) The combining unit 56 determines whether j is M or not. When it is M, it progresses to step S2411, and when that is not right, it progresses to step S2421.

  (Step S2421) The combining unit 56 increments j by 1, and returns to step S2414.

  In the flowchart of FIG. 24, it is determined whether or not two or more attribute value sets satisfy a combining condition in the case of having a unique identifier, and if so, the two or more attribute value sets are combined. Also good.

  Also, in the flowchart of FIG. 24, it is determined whether or not the attribute value set of 3 or more satisfies the combining condition in the case where the unique identifier does not exist, and if so, the attribute value set of 3 or more is combined. It may be.

  FIG. 25 is a flowchart showing the rule acquisition processing in step S2206 of the flowchart of FIG.

  (Step S2501) The rule acquisition unit 57 acquires all attribute value sets stored in the table storage unit 51. Here, it is assumed that M attribute value sets have been acquired.

  (Step S2502) The rule acquisition unit 57 performs crawling and acquires one or more Web pages from the Web. Here, it is assumed that N Web pages have been acquired.

  (Step S2503) The rule acquisition unit 57 sets 1 to the counter i.

  (Step S2504) The rule acquisition unit 57 sets 1 to the counter j.

  (Step S2505) The rule acquisition unit 57 sets 0 to the variable count.

  (Step S2506) The rule acquisition unit 57 determines whether or not a character string having the attribute value set [i] exists in the Web page [j]. When it exists, it progresses to step S2507, and when that is not right, it progresses to step S2511.

  (Step S2507) The rule acquisition unit 57 acquires a character string in which the attribute value set [i] exists from the Web page [j].

  (Step S2508) The rule acquisition unit 57 acquires a rule from the acquired character string.

  (Step S2509) The rule acquisition unit 57 increments count by 1.

  (Step S2510) The rule acquisition unit 57 sets the acquired attribute value set in the variable rules [count].

  (Step S2511) The rule acquisition unit 57 determines whether j is N or not. When it is N, it progresses to step 2513, and when that is not right, it progresses to step S2512.

  (Step S2512) The rule acquisition unit 57 increments j by 1, and returns to step S2506.

  (Step S2513) The rule acquisition unit 57 determines whether the count is 1 or more. When it is 1 or more, the process proceeds to step S2514. Otherwise, the process proceeds to S2515.

  (Step S 2514) The rule acquisition unit 57 accumulates rules in the rule storage unit 52.

  (Step S2515) The rule acquisition unit 57 determines whether i is M or not. If it is M, the process returns to the upper process, and if not, the process proceeds to step S2516.

  In the flowchart of FIG. 25, one or more rules are acquired from N Web pages for each attribute value set. Thereby, the acquired rule can be accumulated in the rule storage unit 52 for each attribute of the attribute value set.

  Further, in the flowchart of FIG. 25, for all M attribute value sets, one or more rules are acquired from N Web pages, and all the acquired rules are collectively stored in the rule storage unit 52. May be.

(Concrete example)
Next, a specific example of the operation of the database construction device 5 will be described.

  (Example 1: Acquisition of attribute value set)

  In this example, it is assumed that the rule storage unit 52 stores the rules shown in FIG. The rule includes an ID (item name: ID) for uniquely identifying each record, a rule (item name: acquisition rule), and an attribute value set attribute (item name: attribute) that can be acquired by the rule. It has correspondence.

  First, the acquisition unit 55 acquires all the rules in FIG.

  Next, it is assumed that the acquisition unit 55 performs crawling, acquires one or more Web pages from the Web, and acquires the Web pages illustrated in FIGS. 27 and 28. The web page in FIG. 27 is a text document, and the web page in FIG. 28 is an HTML document.

  Next, the acquisition unit 55 acquires a character string that matches each of the acquired rules from the acquired Web page. Here, in the text document of FIG. 27, the character string “Co., Ltd.” that matches the rule “$ company name $, $ product name $, $ unique name $. *? Release” of “ID = 011” in FIG. "Released non-defective home appliances, full HD TV, and BeautyScreen" are on the 12th line. In addition, the HTML document of FIG. 28 matches the rule “<h2> $ company name $, $ product name $, and unique name $. *? Release </ h2>” of “ID = 021” in FIG. The character string “<h2> Ryohin Denki Home Appliances Co., Ltd., Full HD TV, Releases BeautyScreen </ h2>” exists on the 12th line. Therefore, the acquisition unit 55 acquires the two character strings.

  Next, the acquisition unit 55 uses the rules “$ company name $, $ product name $,” used when acquiring the character string from the acquired character string “Released Ryohin Denki Co., Ltd., Full HD TV, BeautyScreen”. The character string in the variable part indicating the attribute value of “$ unique name $. *? Release” is acquired. Then, the acquisition unit 55 acquires “Ryosan Co., Ltd.” as the company name, “Full HD TV” as the product name, and “BeautyScreen” as the unique name. In addition, the acquisition unit 55 uses the rule “<h2> $ used when acquiring the character string from the acquired character string“ <h2> Ryohin Denki Home Appliances Co., Ltd., Full HD TV, Releases BeautyScreen </ h2> ”. A character string in a variable portion indicating an attribute value of “company name $, $ product name $, $ unique name $. *? Release </ h2>” is acquired. Then, the acquisition unit 55 acquires “Ryosan Co., Ltd.” as the company name, “Full HD TV” as the product name, and “BeautyScreen” as the unique name.

  Since the two acquired attribute value sets overlap, the acquiring unit 55 normally performs a so-called unique process, and finally ends up with one attribute value set “Ryohin Denki Co., Ltd. | Full HD TV | BeautyScreen”. To get.

  Next, the acquisition unit 55 adds the acquired attribute value set “Ryohin Denki Co., Ltd. | Full HD TV | BeautyScreen” to the second table. The situation at this time is, for example, FIG. At this time, for example, if there is no table having an attribute corresponding to the attribute value set in the table storage unit 51, the acquisition unit 55 first generates a table having an attribute corresponding to the acquired attribute value set. The attribute value set acquired is added to the table.

  Further, for example, assume that the URL of the Web page in FIG. 27 is “http: // official. Beautyscreen.com”. In this case, the acquisition unit 55 may acquire the URL. Then, the acquisition unit 55 may add the acquired URL to the second table in association with the acquired attribute value set. The state at this time is, for example, FIG.

  Further, for example, it is assumed that the web page score management information having the URL “http: // official. Beautyscreen.com” and the score “100” associated with each other is stored in the score storage unit 53. In this case, the acquisition unit 55 acquires the score “100” corresponding to the URL from the score storage unit 53. Then, the acquisition unit 55 adds the acquired score to the second table in association with the acquired attribute value set. The situation at this time is, for example, FIG.

  (Example 2: Generation of attribute value set)

  In this example, it is assumed that the table storage unit 51 stores the second table shown in FIGS. Each of the second tables has two attributes and has attribute values of the attributes. In the second table, the item name “ID” is not usually an attribute value, but may be an attribute value. In the second table of FIGS. 32 and 33, “*” is a unique identifier, which indicates that the attribute to which the unique identifier is attached is unique.

  First, the combining unit 56 acquires the second table in FIG. 32 and FIG. 33 having the unique identifier.

  Next, the combining unit 56 sets the attribute value set of “ID = 014” in FIG. 33 which is an attribute value set having the same attribute and attribute value as the unique name “PAT-Analyzer” of “ID = 011” in FIG. Acquire “PAT-Analyzer | software”. Then, the combining unit 56 sets the attribute value set “PAT-Analyzer” of “ID = 011” in FIG. 32 and the attribute value set “PAT-Analyzer | software” of “ID = 014” in FIG. And a new attribute value set “Patent Analysis Co., Ltd. | PAT-Analyzer | Software” is generated.

  Next, the combining unit 56 accumulates the generated attribute value set “Patent Analysis Co., Ltd. | PAT-Analyzer | software” in the table storage unit 51 as the attribute value set of the first table. At this time, the combining unit 56 generates an attribute value set in the form of a table as shown in FIG. At this time, for example, if there is no table having an attribute corresponding to the attribute value set in the table storage unit 51, the combining unit 56 first generates a table having an attribute corresponding to the generated attribute value set. The attribute value set acquired is added to the table.

  Next, in the same manner as described above, the combining unit 56 sets “ID = 011” in FIG. 33, which is an attribute value set having the same attribute and attribute value as the unique name “BeautyScreen” of “ID = 012” in FIG. The attribute value set “BeautyScreen | Full HD TV” is acquired. Then, the combining unit 56 sets the attribute value set “ROKYO ELECTRIC CO., LTD. | BeautyScreen” of “ID = 012” in FIG. 32 and the attribute value set “BeautyScreen | Combined to generate a new attribute value set “Ryohin Denki Co., Ltd. | BeautyScreen | Full HD TV”. Then, the combining unit 56 accumulates the generated attribute value set “non-defective household appliances | BeautyScreen | full HD TV” in the table storage unit 51 as described above.

  Note that the combining unit 56 does not have an attribute value set having the same attribute and attribute value in the attribute value set of FIG. 33 for the attribute value sets of “ID = 013” and “ID = 014” in FIG. Therefore, the process of combining is not performed.

  Next, the combining unit 56 acquires the second table in FIG. 34, FIG. 35, and FIG. 36 that does not have a unique identifier.

  Next, the combining unit 56 acquires one attribute value set from each of the acquired second tables, and generates a combination of three attribute value sets. For example, the attribute value set of “ID = 011” in FIG. 34, the attribute value set of “ID = 011” in FIG. 35, and the attribute value set of “ID = 011” in FIG. 36 are acquired, and these three attribute value sets are acquired. It is assumed that the combination of ““ Patent Analysis Co., Ltd. | Information Service ”,“ Electrical Machinery Manufacturing | 760,000 Yen ”,“ Building Co., Ltd. | 640,000 Yen ”” is generated. Since the combination does not satisfy the coupling condition, the coupling unit 56 does nothing.

  Next, in the same way as described above, the combining unit 56 generates an attribute value set combination in order to generate an attribute value set “ID = 013” in FIG. The attribute value set “ID = 011” “general construction | 640,000,000 yen” and the attribute value set “ID = 011” in FIG. 36 “Building Co., Ltd. | 640,000,000 yen” are acquired, and the attribute value set It is assumed that the combination ““ Building Co., Ltd. | General Construction ”,“ General Construction | 640,000 Yen ”,“ Building Co., Ltd. | 640,000 Yen ”” is generated. Since the combination satisfies the combination condition, the combining unit 56 combines the attribute value sets to generate a new attribute value set “Building Co., Ltd. | General Construction | 640,000,000 Yen”. Then, the combining unit 56 accumulates the generated attribute value set in the table storage unit 51 in the same manner as described above.

  (Example 3: Rule acquisition)

  In this example, it is assumed that the table storage unit 51 stores the first table generated in Example 2 and the second table shown in FIGS.

  First, it is assumed that the rule acquisition unit 57 performs crawling, acquires one or more Web pages from the Web, and acquires the Web pages shown in FIGS.

  Next, the rule acquisition unit 57 acquires one or more rules in which the attribute value appears from the acquired Web pages of FIGS. 27 and 28, using the acquired attribute value set in order. Here, for example, it is assumed that the rule acquisition unit 57 uses the attribute value set “good product home appliances | BeautyScreen” of “ID = 012” in FIG. Then, the rule acquisition unit 57 acquires a character string in which both of the two attribute values that the attribute value set has exist from the Web pages of FIGS. 27 and 28. In the text document of FIG. 27, a character string having the attribute value set exists on the 12th and 13th to 14th lines. In the HTML document of FIG. 28, a character string having the attribute value set exists from the 12th line and from the 13th line to the 14th line. Therefore, the rule acquisition unit 57 releases the character strings “Ryohin Denki Home Appliances Co., Ltd., Full HD TV, Beauty Screen” and “Ryohin Denki Co., Ltd. Announced to be released on October 14, 2011 "," <h2> Ryohin Denki Home Appliances Co., Ltd. launches Full HD TV, BeautyScreen </ h2> "," <p> Announced to release HD TV and BeautyScreen on October 14, 2011.

  Here, the variable corresponding to the company name is “% company name%”, and the variable corresponding to the unique name is “% unique name”. Next, the rule acquisition unit 57 replaces the attribute value in the acquired character string with these variables, and releases the new rules “% company name%, full HD TV,% unique name%”, “% company Name% announced that it will launch a new full HD TV,% proprietary name% on October 14, 2011, "<h2>% company name%, full HD TV,% unique "Name% released </ h2>", "<p>% Company name% announced that it will release a new full HD TV,% proprietary name%, on October 14, 2011 < / P> ".

  It should be noted that “Full HD TV” included in the acquired rule is a product name from the second table of FIG. Further, it is clear that “October 14, 2011” in the acquired rule is a date from the format. Accordingly, the rule acquisition unit 57 replaces these product names and dates with the variables corresponding to these rules “Release% company name%,% product name%,% unique name%”, “% company name% is , "% Brand name%,% unique name% announced to be released on <new line>% date%", "<h2>% company name%,% product name%,% unique name% released </ H2> ”,“ <p>% company name% announced that it will release% new product name% and% unique name% on <new line>% date% </ p> ” May be.

  Next, the rule acquisition unit 57 accumulates the acquired rules in the rule storage unit 52.

  As mentioned above, according to the database construction apparatus 5 by this Embodiment, all the information on Web can be acquired and it can be made into a database.

  Further, according to the database construction device 5 according to the present embodiment, the database information is automatically combined, and an attribute value set having a larger number of attributes than the number of attributes is generated from an attribute value set having a smaller number of attributes. be able to.

  Moreover, according to the database construction device 5 according to the present embodiment, it is possible to automatically increase the rules for acquiring the attribute value set.

(Embodiment 4)
The first embodiment will be described more clearly. FIG. 1 is a block diagram of a database construction device 1 according to the present embodiment. The database construction device 1 includes a company database 101, a product database 102, a unique name database 103, a company product database 104, a company unique name database 105, a product unique name database 106, a company product unique name database 107, a company rule database 108, a product rule. A database 109, a unique name rule database 110, a knowledge multiplication rule database 111, a company name acquisition unit 112, a product name acquisition unit 113, a unique name acquisition unit 114, a knowledge acquisition unit 115, and a rule learning unit 116 are provided.

  The company database 101 can store one or more company names. The company name is a name having a legal personality such as a stock company, a limited liability company, a public interest corporation, an incorporated association or an independent administrative corporation that provides goods or services, and an individual name.

  Further, the company name may be information for specifying the company name. The company database 101 may store not only the company name but also information on the document from which the company name has been acquired, the number of acquisitions of the company name, the acquisition rate, and the like. The information for identifying the company of the company name may be any information that can uniquely identify the company. The information specifying the company of the company name may be, for example, information indicating a homepage address, an address, a telephone number, or the like. The information of the document from which the company name has been acquired is, for example, the name of the document, the number of pages, information indicating the URL of the Web site, or the text before and after the company name, or the HTML tag structure.

  Storage is a concept that includes long-term storage using a non-volatile recording medium and temporary storage using a volatile recording medium. Hereinafter, the concept of storage in the present invention is the same.

  Further, the process of storing the company name in the company database 101 does not matter. For example, a company name may be stored in the company database 101 via a recording medium, or a company name transmitted via a communication line or the like may be stored in the company database 101. Alternatively, the company name input via the input device may be stored in the company database 101. The company database 101 is preferably a non-volatile recording medium, but can also be realized by a volatile recording medium.

  A product database 102, a unique name database 103, a company product database 104, a company unique name database 105, a product unique name database 106, a company product unique name database 107, a company rule database 108, a product rule database 109, a unique name rule, which will be described later. Information is stored in each component of the database 110, the knowledge multiplication rule database 111, the product similar group code correspondence database 21, the company similar group code unique name database 22, the database 31, the database 401, and the product code management means 4021. Since the process and the means for realizing them are the same as those of the company database 101, description thereof is omitted.

  The product database 102 can store one or more product names. The product name is the name of a product or service sold or provided by a corporation and an individual. The product names are specifically “chocolate”, “computer”, “delivery”, “moving service”, and the like. The names of goods and services mentioned here are not necessarily the names of goods and services described in the classification of goods and services (hereinafter referred to as similar group code table) stipulated in Article 6 of the Enforcement Regulations of the Trademark Law. It does not have to comply.

  Further, the product name may be information for specifying the product name. The product database 102 may store not only the product name but also the information of the document from which the product name is acquired, the number of acquisitions of the product name, the acquisition ratio, and the like. The information specifying the product with the product name may be anything as long as it is information that can uniquely specify the product. The information specifying the product with the product name may be, for example, information indicating a homepage address, an address, a telephone number, or the like. The information of the document from which the product name has been acquired is, for example, the name of the document, the number of pages, information indicating the URL of the Web site, or the text before and after the product name, the HTML tag structure, or the like.

  The unique name database 103 can store unique names of one or more products. The unique name of the product is a name that can identify the product and service. Specific names for identifying the product and service include “Yukimi Daifuku (registered trademark)”, “Takkyubin (registered trademark)”, “PatentSearch Assistant (registered trademark)”, and the like. In addition, the proper name of the product mentioned here may be a trademark registered as a trademark, or may not be.

  Further, the unique name of the product may be information for specifying the unique name of the product. The unique name database 103 may store not only the unique name of the product but also the information of the document that acquired the unique name of the product, the number of acquisitions of the unique name of the product, the acquisition ratio, and the like. The information that identifies the unique name of the product may be any information that can uniquely identify the unique name of the product. Information that can uniquely identify a unique name of a product may be, for example, a trademark application number or a publication number. The information of the document from which the product unique name is acquired is, for example, the document name, the number of pages, information indicating the URL of the Web site, or the text before and after the product unique name, the HTML tag structure, or the like.

  The company product database 104 may store one or more pairs having a company name and a product name.

  The company product database 104 stores not only the company name and the product name but also the information of the document that acquired the company name and product name in the same record, or the number of times the company name and product name are acquired in the same record, the acquisition frequency, Occurrence frequency, the average distance based on the number of characters, the number of morphemes, or the number of words listed between the company name and the product name in the document, or the total or average of the number of acquisitions and appearance frequency values of each rule described later The value calculated by using any method may be used. The information of the document in which the company name and product name in the same record are acquired is, for example, the document name and the number of pages, or information indicating the URL of the website, the text before and after the company name and product name, and the HTML tag Such as structure.

  The company unique name database 105 may store one or more sets having a company name and a product unique name.

  The company unique name database 105 stores not only the company name and the unique name of the product, but also the information of the document that acquired the company name and the unique name of the product in the same record, or the company name and the unique name of the product in the same record. Number of acquisitions, acquisition frequency and co-occurrence frequency, the average number of characters, the number of morphemes, the number of words, etc. written between the company name and the unique name of the product in the document, or the acquisition number of each rule described later And a value calculated by using some method such as a sum or average of appearance frequency values may be stored. The information of the document in which the company name and the unique name of the product in the same record are acquired is, for example, the name of the document, the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  The product unique name database 106 may store one or more sets having product names and product unique names.

  The company unique name database 105 stores not only the product name and the product unique name, but also the information of the document in which the product name and product unique name in the same record are acquired, or the product name and product unique name in the same record. Acquisition count, acquisition frequency, co-occurrence frequency, average distance between the product name and the product unique name in the document, the number of characters, the number of morphemes, the number of words, etc. And a value calculated by using some method such as a sum or average of appearance frequency values may be stored. The information of the document in which the product name and the unique name of the product in the same record are acquired is, for example, the document name and the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  The company product unique name database 107 may store one or more sets having a company name, a product name, and a product unique name.

  The company product unique name database 107 includes not only the company name, the product name, and the product unique name, but also the information on the document that acquired the company name, product name, and product unique name in the same record, or the same record. Number of acquisitions of company name, product name and unique name of product, acquisition frequency and co-occurrence frequency, number of characters written between each of company name, product name and unique name of product in document, morpheme number, or An average distance based on the number of words or the like, or a value calculated by using some method such as a total or average of the number of acquired rules and the appearance frequency described later may be stored. The information of the document in which the product name and the unique name of the product in the same record are acquired is, for example, the document name and the number of pages, information indicating the URL of the website, the text before and after the unique name of the product, or HTML For example, tag structure.

  In addition, the company product unique name database 107 may store trademark information that has already been published. In that case, the company product unique name database 107 may store information such as trademark identification information (for example, registration number) and registration status.

  The company rule database 108 may store one or more company rules that are rules for obtaining a company name from a document.

  A document is a collection of one or more sentences. The company name, product name, and product unique name do not necessarily have to be written, and are not only Japanese but also English, Chinese, and all other languages. Alternatively, two or more languages may be mixed. For example, the document may be a Web page, HTML, XML, Word file, or the like.

  Obtaining a company name from a document means that a part in which the company name is described is determined from the document and the company name is extracted.

  The company rule is a rule for extracting a company name from a document. The company rule may be, for example, a sentence pattern in which the company name can be directly extracted, or a sentence pattern that specifies a sentence in which the company name is described. If the document is a Web page, the company rule may be an HTML tag structure pattern that directly extracts the company name, or an HTML tag structure pattern that specifies a sentence in which the company name is described. There may be. Moreover, the same pattern as the above which cannot extract a company name may be sufficient. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the company rules may be combined with the above-described rules. Specifically, the company rule is “<td> company name </ td> <td> company% company name% <” in the case of a company rule for acquiring a company name described in the “% company name%” portion. / Td> ”and the like.

  The company rule database 108 may store not only the company rules but also the number of acquisitions and the acquisition ratio of the company rules, or the targets extracted by the company rules. The target to be extracted by the company rules is when there is a possibility of acquiring multiple types of information, such as when acquiring a company name or when acquiring a sentence including the company name as a result of using the company rule from a document , A value that clearly specifies what to get. Specifically, the company rule database 108 may store values such as “company name” or “text”.

  The merchandise rule database 109 can store one or more merchandise rules that are rules for obtaining a merchandise name from a document.

  Obtaining the product name from the document means that the part in which the product name is described is determined from the document and the product name is extracted.

  The product rule is a rule for acquiring a product name from a document. For example, the product rule may be a text pattern from which a product name can be directly extracted, or a text pattern that specifies a text in which the product name is described. When the document is a Web page, the product rule may be an HTML tag structure pattern that directly extracts the product name, or an HTML tag structure pattern that specifies a sentence in which the product name is described. There may be. Further, the product rule may be a pattern similar to the above in which the product name cannot be extracted. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the product rules may be a combination of the above-described rules. Specifically, in the case of a product rule for acquiring a product name described in the “% product name%” part, the product rule is “<title> product name list </ title>. *? <Div. *? > <H [1-5]. *?> List <h [1-5]> </ div> <ul. *?> (<Li. *?> <A. *?>% Product name% <// a> </ </ li>) + </ ul> ".

  The product rule database 109 may store not only the product rules, but also the number of acquisitions and the acquisition ratio of the product rules, or the targets extracted by the product rules. The target of product rule extraction is when there is a possibility of acquiring multiple types of information, such as when acquiring a product name as a result of using a product rule from a document, or when acquiring a sentence containing the product name , A value that clearly specifies what to get. Specifically, the product rule database 109 may store values such as “product name” or “text”.

  The unique name rule database 110 can store one or more unique name rules, which are rules for acquiring a unique name of a product from a document.

  Obtaining the unique name of the product from the document means that the part in which the unique name of the product is described is determined from the document and the product name is extracted.

  The unique name rule is a rule for acquiring a unique name of a product from a document. For example, the unique name rule may be a sentence pattern that can directly extract the unique name of the product, or may be a sentence pattern that specifies a sentence in which the unique name of the product is described. If the document is a Web page, the unique name rule may be an HTML tag structure pattern that directly extracts the unique name of the product, and the HTML rule that specifies a sentence in which the unique name of the product is described. It may be a tag structure pattern. The unique name rule may be a pattern similar to the above in which the unique name of the product cannot be extracted. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the unique name rule may be a combination of the above-described rules. Specifically, in the case of the unique name rule for acquiring the unique name of the product described in the “% unique name%” part, the unique name rule is “<title> product list </ title>. *? <Span”. ... ?? list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> .

  The unique name rule database 110 may store not only the unique name rule but also the number of acquisitions and the acquisition ratio of the unique name rule, or the target extracted by the unique name rule. The target to be extracted by the unique name rule is the acquisition of multiple types of information, such as when acquiring a unique name of a product or when acquiring a sentence containing the unique name of a product as a result of using the unique name rule from a document. If there is a possibility, it is a value that specifies what to get. Specifically, the unique name rule database 110 may store values such as “unique name of product” or “text”.

  The knowledge proliferation rule database 111 can store one or more knowledge proliferation rules that are rules for acquiring information with a larger number of components from a document using information with a smaller number of components.

The acquisition of information with a larger number of components using information with a smaller number of components is as follows.
(1) Using information stored in the company database 101 to acquire information stored in the company product database 104, the company unique name database 105, and the company product unique name database 107. (2) Product database Obtain information stored in the company product database 104, the product unique name database 106, and the company product unique name database 107 using the information stored in the information 102. (3) Stored in the unique name database 103 Information stored in the company unique name database 105, the product unique name database 106, and the company product unique name database 107 (4) information stored in the company product database 104 Use the meeting Obtaining information stored in the product unique name database 107 (5) Obtaining information stored in the company unique name database 107 using information stored in the company unique name database 105 ( 6) Acquiring information stored in the company product unique name database 107 using the information stored in the product unique name database 106

  The knowledge proliferation rule is a rule for acquiring the remaining two or one type of information using one or two types of information from a company name, a product name, or a unique name of a product from a document. Specifically, the knowledge proliferation rule includes the first knowledge proliferation rule, the second knowledge proliferation rule, the third knowledge proliferation rule, the fourth knowledge proliferation rule, the fifth knowledge proliferation rule, and the sixth knowledge proliferation rule. And a seventh knowledge proliferation rule, an eighth knowledge proliferation rule, a ninth knowledge proliferation rule, a tenth knowledge proliferation rule, an eleventh knowledge proliferation rule, and a twelfth knowledge proliferation rule.

  The first knowledge proliferation rule is a rule for acquiring a combination of a company name and a product name from a document using a company name. Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The first knowledge proliferation rule may be, for example, a sentence pattern in which a product name related to a company name can be directly extracted from a document in which the company name is described. It may be a sentence pattern that identifies the existing sentence. If the document is a Web page, the first knowledge proliferation rule may be an HTML tag structure pattern that directly extracts a product name related to the company name from the Web page on which the company name is described. The pattern of the HTML tag structure which specifies the sentence with which the brand name is described may be sufficient. Further, the first knowledge proliferation rule may be a pattern similar to the above in which a product name related to a company name cannot be extracted from documents in which the company name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the first knowledge proliferation rule may be a combination of the above-described rules. Specifically, when it is shown that the combination of the company name described in the “% company name%” part and the product name described in the “% product name%” part is acquired, the first knowledge proliferation rule is For example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A . *?>% Product name% </a> </ li>) + </ ul> ".

  The second knowledge proliferation rule is a rule for acquiring a combination of a company name and a product unique name from a document using the company name.

  Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The second knowledge proliferation rule may be, for example, a sentence pattern in which a proper noun of a product related to a company name can be directly extracted from a document in which the company name is described. It may be a sentence pattern that identifies sentences in which nouns are written. When the document is a Web page, the second knowledge proliferation rule is a pattern of an HTML tag structure that directly extracts a proper noun of a product related to the company name from the Web page in which the company name is described. Alternatively, it may be an HTML tag structure pattern that identifies a sentence in which a proper noun of a product is described. Further, the second knowledge proliferation rule may be a pattern similar to the above in which a proper noun of a product related to the company name cannot be extracted from the document in which the company name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the second knowledge proliferation rule may be a combination of the above-described rules. Specifically, when it is shown that the combination of the company name described in the “% company name%” part and the unique name of the product described in the “% unique name%” part is acquired, the second knowledge multiplication The rule is, for example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?> Product list </ span> <ul. *?> (<Li. *? > <A. *?>% Proper name% </a> </ li>) + </ ul> ".

  The third knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a product unique name from a document using the company name.

  Using the company name may be, for example, selecting a document in which one or more company names stored in the company database 101 are described from the documents described above, and the company name is described. For example, a surrounding character string may be extracted.

  The third knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a product name related to a company name and a proper noun of the product among documents in which the company name is described, It may be a sentence pattern that directly extracts the product name and identifies a sentence that includes the unique name of the product, or a sentence pattern that identifies a sentence that includes the product name and directly extracts the proper noun of the product. And a sentence pattern that identifies a sentence in which the proper noun of the product is described. If the document is a Web page, the third knowledge proliferation rule has an HTML tag structure that directly extracts the product name and the proper noun of the product from the Web page on which the company name is described. It may be a pattern, or it may be an HTML tag structure pattern that directly identifies a product name and identifies a sentence that includes the unique name of the commodity, identifies a sentence that includes the product name, and directly identifies a proper noun of the commodity It may be an HTML tag structure pattern, or an HTML tag structure pattern that identifies a sentence in which a product name and a proper noun of the product are described. Further, the third knowledge proliferation rule may be a pattern similar to the above in which the product name related to the company name and the proper noun of the product cannot be extracted from the document in which the company name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the third knowledge proliferation rule may be a combination of the above-mentioned rules. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. In the case of indicating that a set of unique names is acquired, the third knowledge proliferation rule is, for example, “<title>% company name% (handling)? Merchandise </ title>. *? <Span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ ul>

  The fourth knowledge proliferation rule is a rule for acquiring a combination of a company name and a product name from a document using a product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The fourth knowledge proliferation rule may be, for example, a sentence pattern in which a company name related to a product name can be directly extracted from a document in which the product name is described, and the company name is described. It may be a sentence pattern that identifies the existing sentence. If the document is a Web page, the fourth knowledge proliferation rule is a pattern of an HTML tag structure that directly extracts a company name related to a product name from Web pages on which the product name is described. Alternatively, it may be a pattern of an HTML tag structure that specifies a sentence in which a company name is described. Further, the fourth knowledge proliferation rule may be a pattern similar to the above in which the company name related to the product name cannot be extracted from the document in which the product name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the fourth knowledge proliferation rule may be a combination of the above-described rules. Specifically, when indicating that the combination of the company name described in the “% company name%” part and the product name described in the “% product name%” part is to be acquired, the fourth knowledge proliferation rule is For example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A . *?>% Product name% </a> </ li>) + </ ul> ".

  The fifth knowledge proliferation rule is a rule for acquiring a combination of a product name and a product unique name from a document using the product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The fifth knowledge proliferation rule may be, for example, a sentence pattern in which a unique name of a product related to the product name can be directly extracted from a document in which the product name is described. It may be a sentence pattern that identifies a sentence in which a name is described. When the document is a Web page, the fifth knowledge proliferation rule is an HTML tag structure pattern that directly extracts a unique name of a product related to the product name from the Web page on which the product name is described. It may be a pattern of an HTML tag structure that specifies a sentence in which a unique name of a product is described. Further, the fifth knowledge proliferation rule may be a pattern similar to the above in which the unique name of the product related to the product name cannot be extracted from the document in which the product name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the fifth knowledge proliferation rule may be a combination of the above-described rules. Specifically, if it is shown that the combination of the product name described in the “% product name%” part and the product name described in the “% unique name%” part is acquired, The rule is, for example, “<span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> <// li>) + </ ul> ".

  The sixth knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a product unique name from a document using the product name.

  Using the product name may be, for example, selecting a document in which one or more product names stored in the product database 102 are described from the documents described above, and the product name is described. For example, a surrounding character string may be extracted.

  The sixth knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a company name related to a product name and a proper noun of the product among documents in which the product name is described, It may be a sentence pattern that directly extracts the company name and identifies the sentence including the unique name of the product, or may be a sentence pattern that identifies the sentence including the company name and directly extracts the proper noun of the product. And a sentence pattern that identifies a sentence in which the proper noun of the product is described. Further, when the document is a Web page, the sixth knowledge proliferation rule is an HTML tag structure that directly extracts a company name related to a product name and a proper noun of the product from the Web page on which the product name is described. It may be a pattern of HTML tag structure that directly identifies the company name and identifies the sentence including the unique name of the product, and identifies the sentence including the company name and directly identifies the proper noun of the commodity. The pattern of the HTML tag structure to identify may be sufficient, and the pattern of the tag HTML tag structure which identifies the sentence in which the company name and the proper noun of the product are described may be used. Further, the sixth knowledge proliferation rule may be a pattern similar to the above in which a company name and a proper noun of a product related to a product name cannot be extracted from a document in which the product name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the sixth knowledge proliferation rule may be a combination of the above-described rules. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. In the case of indicating that a set of unique names is acquired, the sixth knowledge proliferation rule is, for example, “<title>% company name% (handling)? Merchandise </ title>. *? <Span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ ul>

  The seventh knowledge proliferation rule is a rule for acquiring a combination of a company name and a product unique name from a document using a product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The seventh knowledge proliferation rule may be, for example, a sentence pattern that can directly extract a company name related to a unique name of a product from documents in which the unique name of the product is described. It may be a sentence pattern that identifies sentences in which names are described. If the document is a Web page, the seventh knowledge proliferation rule has an HTML tag structure that directly extracts the company name related to the unique name of the product from the Web page in which the unique name of the product is described. A pattern may be sufficient and the pattern of the HTML tag structure which specifies the text in which the company name is described may be sufficient. Further, the seventh knowledge proliferation rule may be a pattern similar to the above in which a company name related to a unique name of a product cannot be extracted from a document in which the unique name of the product is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the seventh knowledge proliferation rule may be a combination of the above-mentioned rules. Specifically, if you indicate that you want to acquire a combination of the company name listed in the “% company name%” part and the unique name of the product listed in the “% unique name%” part, The rule is, for example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?> Product list </ span> <ul. *?> (<Li. *? > <A. *?>% Proper name% </a> </ li>) + </ ul> ".

  The eighth knowledge proliferation rule is a rule for acquiring a combination of a product name and a product unique name from a document using the product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The eighth knowledge proliferation rule may be, for example, a sentence pattern that can directly extract the product name related to the product unique name from the document in which the product unique name is described. It may be a sentence pattern that specifies a sentence in which a name is described. Further, when the document is a Web page, the eighth knowledge proliferation rule has an HTML tag structure for directly extracting a product name related to the product unique name from the Web page in which the product unique name is described. A pattern may be sufficient and the pattern of the HTML tag structure which specifies the text in which the brand name is described may be sufficient. In addition, the eighth knowledge proliferation rule may be a pattern similar to the above in which a product name related to a product unique name cannot be extracted from a document in which a product unique name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the eighth knowledge proliferation rule may be a combination of the above-described rules. Specifically, if it is indicated that the combination of the product name described in the “% product name%” part and the product unique name described in the “% unique name%” part is acquired, The rule is, for example, “<span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> <// li>) + </ ul> ".

  The ninth knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a product unique name from a document using the product unique name.

  Using the unique name may be, for example, selecting a document in which the unique names of one or more products stored in the unique name database 103 are described from the above-described documents. It is also possible to extract a surrounding character string in which is described.

  The ninth knowledge proliferation rule is, for example, a sentence pattern in which a company name and a product name related to a product unique name can be directly extracted from a document in which a product unique name is described. It may be a sentence pattern that directly extracts a company name and identifies a sentence including the product name, or a sentence pattern that identifies a sentence including the company name and directly extracts the product name. It may be a sentence pattern that specifies a sentence in which a name is described. Also, if the document is a Web page, the ninth knowledge proliferation rule is an HTML code that directly extracts the company name and product name related to the product unique name from the Web page on which the product unique name is described. It may be a tag structure pattern, or it may be an HTML tag structure pattern that directly identifies the company name and identifies the text including the product name, and identifies the text including the company name and directly identifies the product name. It may be an HTML tag structure pattern, or an HTML tag structure pattern that specifies a sentence in which a company name and a product name are described. Further, the ninth knowledge proliferation rule may be a pattern similar to the above in which the company name and the product name related to the product unique name cannot be extracted from the document in which the product unique name is described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Further, the ninth knowledge proliferation rule may be a combination of the above-described rules. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. In the case of indicating that a set of unique names is acquired, the ninth knowledge proliferation rule is, for example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ ul>

  The tenth knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a unique name of a product from a document by using the combination of the company name and the product name.

  Using a combination of a company name and a product name includes, for example, one or more company names stored in the company database 101 and one or more product names stored in the product database 102 from the document described above. It is also possible to select a document that has a name of a company and a character string around the product name.

  The tenth knowledge proliferation rule is, for example, a sentence pattern in which a unique name of a product related to a company name and a product name can be directly extracted from documents in which the company name and the product name are described. Alternatively, it may be a sentence pattern that specifies a sentence in which the unique name of the product is described. If the document is a Web page, the tenth knowledge proliferation rule is an HTML that directly extracts a unique name of a product related to the company name and the product name from the Web page in which the company name and the product name are described. The tag structure pattern may be an HTML tag structure pattern that identifies a sentence in which a product unique name is described. Further, the tenth knowledge proliferation rule may be a pattern similar to the above in which a unique name of a product related to the company name and the product name cannot be extracted from the document in which the company name and the product name are described. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the tenth knowledge proliferation rule may be a combination of the above-mentioned rules. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. 10th knowledge proliferation rule, for example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ ul>

  The eleventh knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a unique name from a document using a combination of the company name and the unique name of the product.

  Using a combination of company name and unique name means, for example, one or more company names stored in the company database 101 from the above-mentioned document and one or more unique names of one or more products stored in the unique name database 103 May be selected, or a peripheral character string in which the company name and the unique name of the product are described may be extracted.

  The eleventh knowledge proliferation rule is, for example, a sentence pattern in which a product name related to a company name and a product unique name can be directly extracted from a document in which the company name and product unique name are described. It may be a sentence pattern that identifies a sentence in which a product name is described. When the document is a Web page, the eleventh knowledge proliferation rule directly selects a product name related to the company name and the unique name of the product among the Web pages on which the company name and the unique name of the product are described. The pattern of the HTML tag structure to extract may be sufficient, and the pattern of the HTML tag structure which specifies the text in which the brand name is described may be sufficient. The eleventh knowledge proliferation rule is a pattern similar to the above in which the product name related to the company name and the product unique name cannot be extracted from the document in which the company name and the product unique name are described. Also good. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the eleventh knowledge proliferation rule may be a combination of the above-mentioned rules. Specifically, the eleventh knowledge proliferation rule consists of the company name described in the “% company name%” part, the product name described in the “% product name%” part, and the “% unique name%”. In the case where it indicates that a set of unique names of products described in the portion of “<title>%” is to be acquired, for example, “<title>% company name% (handling)? Product </ title>. *? <Span. *?>% Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ Ul> ".

  The twelfth knowledge proliferation rule is a rule for acquiring a combination of a company name, a product name, and a unique name from a document using a combination of a product name and a unique name of a product.

  Using a combination of a product name and a unique name means, for example, one or more product names stored in the product database 102 from the above-mentioned document and a unique name of one or more products stored in the unique name database 103. May be selected, a product name and a surrounding character string in which a product unique name is described may be extracted.

  The twelfth knowledge proliferation rule is, for example, a sentence pattern that can directly extract a company name related to a product name and a product unique name from a document in which a product name and a product unique name are described. It may be a sentence pattern that specifies a sentence in which a company name is described. Further, when the document is a Web page, the twelfth knowledge proliferation rule directly selects the company name related to the product name and the product unique name from the Web page on which the product name and the product unique name are described. The pattern of the HTML tag structure to extract may be sufficient, and the pattern of the HTML tag structure which specifies the text in which the company name is described may be sufficient. The twelfth knowledge proliferation rule is a pattern similar to the above in which the company name related to the product name and the product unique name cannot be extracted from the document in which the product name and the product unique name are described. Also good. Each pattern may be a fixed character string, a regular expression, or a part of speech or the number of characters. Furthermore, the twelfth knowledge proliferation rule may be a combination of the above-mentioned rules. Specifically, the company name listed in the “% company name%” part, the product name listed in the “% product name%” part, and the product listed in the “% unique name%” part. In the case of indicating that a set of unique names is acquired, the twelfth knowledge proliferation rule is, for example, “<title>% company name% (handling)? Merchandise </ title>. *? <Span. *?> % Product name% list </ span> <ul. *?> (<Li. *?> <A. *?>% Proper name% </a> </ li>) + </ ul> .

  The knowledge proliferation rule database 111 may store not only the knowledge proliferation rules but also the types of knowledge proliferation rules, the number of acquisitions of the knowledge proliferation rules, the acquisition ratio, or the objects extracted by the knowledge proliferation rules. Targets extracted by knowledge proliferation rules are the results of using knowledge proliferation rules, when acquiring names such as company names, product names, or product unique names from documents, and obtaining sentences that include product unique names. When there is a possibility of acquiring multiple types of information, such as when doing, it is a value that specifies what to acquire. Specifically, the knowledge proliferation rule database 111 may store values such as “name” or “sentence”.

  The knowledge proliferation rule database 111 may hold each of the above knowledge proliferation rules as a flag, or may be a separate database for each of the above knowledge proliferation rules.

  The company name acquisition unit 112 acquires one or more company names from one or more Web server devices using one or more company rules, and accumulates them in the company database 101.

  Specifically, the Web server device is a server device that publishes a Web service or a database. The Web service is not limited to publishing what is configured in a markup language such as HTML format. For example, a system of application software operable on a Web browser or a dedicated application execution environment May be published. The Web server device may be published on the Internet or may be published on a local area.

  The company name acquisition unit 112 acquires a company name from the Web page acquired from the Web server device using a company rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  The acquisition of a company name using a company rule means that a company name registered in the company rule database 108 is applied to a document acquired from a Web server device or an HTML structure, and the company name matches the company rule. Is to get. For example, the company name acquisition unit 112 acquires a character string that applies to a portion described as “% company name%” in an arbitrary company rule.

  When the company name is stored in the company database 101, the company name acquisition unit 112 sets a certain threshold for the number of acquisitions and the acquisition ratio of the used company rules, and stores only information that exceeds the threshold in the company database 101. You may make it register.

  The product name acquisition unit 113 acquires one or more product names from one or more Web server devices using one or more product rules, and accumulates them in the product database 102.

  The product name acquisition unit 113 acquires the product name from the Web page acquired from the Web server device using the product rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  The acquisition of the product name using the product rule means that the product rule registered in the product rule database 109 is applied to the document acquired from the Web server device or the structure of HTML, and the product name that matches the product rule. Is to get. For example, the product name acquisition unit 113 acquires a character string that applies to a portion described as “% product name%” in an arbitrary product rule.

  When the product name acquisition unit 113 accumulates product names in the product database 102, the product name acquisition unit 113 sets a certain threshold for the number of acquisitions and the acquisition ratio of the used product rules, and stores only information that exceeds the threshold in the product database 102. You may make it register.

  The unique name acquisition unit 114 acquires one or more unique names from one or more Web server devices using one or more unique name rules, and accumulates them in the unique name database 103.

  The unique name acquisition unit 114 acquires the unique name of the product from the Web page acquired from the Web server device using the unique name rule. Acquiring a Web page from the Web server device is the same as the processing acquired by a so-called Web browser.

  The acquisition of the unique name of the product using the unique name rule means that the unique name rule stored in the unique name rule database 110 is applied to the document acquired from the Web server device or the structure of HTML, and the like. It is to acquire the unique name of the product that matches the rule. For example, the unique name acquisition unit 114 acquires a character string that applies to a portion described as “% unique name%” in an arbitrary unique name rule.

  When the unique name acquisition unit 114 stores the unique name of the product in the unique name database 103, the unique name acquisition unit 114 sets a certain threshold for the number of acquisitions and the acquisition ratio of the unique name rule used, and only information that exceeds the threshold is obtained. It may be registered in the unique name database 103.

The knowledge acquisition unit 115 acquires predetermined information using predetermined information from Web pages acquired from one or more Web server devices, and accumulates the acquired information in a corresponding database. Specifically, the knowledge acquisition unit 115 acquires predetermined information as follows.
(1) One or more company names and product names, or one or more company names and unique names, or one or more using one or more company names in the company database 101 and one or more knowledge proliferation rules. Get company name, product name and unique name.
(2) One or more company names and product names, or one or more product names and unique names, or one or more using one or more product names in the product database 102 and one or more knowledge proliferation rules. Get company name, product name and unique name.
(3) One or more company names and unique names, or one or more product names and unique names, or 1 using one or more unique names in the unique name database 103 and one or more knowledge proliferation rules. The above company name, product name, and unique name are acquired.
(4) One or more company names, product names, and unique names are acquired using one or more pairs of one or more company names and product names in the company product database 104 and one or more knowledge proliferation rules.
(5) Acquire one or more company names, product names, and unique names using one or more pairs of one or more company names and product unique names in the company unique name database 105 and one or more knowledge proliferation rules. To do.
(6) One or more company names, product names, and unique names are acquired using a combination of one or more product names and product unique names in the product unique name database 106 and one or more knowledge proliferation rules. To do.

  (1) acquires a document including the company name stored in the company database 101 from the Web server device, and stores the first knowledge proliferation rule stored in the knowledge proliferation database for the document, or the first Applying the two-knowledge proliferation rule or the third-knowledge proliferation rule to obtain two or more sets of company name, product name, or product unique name. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  (2) acquires a document including the product name stored in the product database 102 from the Web server device, and stores the fourth knowledge proliferation rule stored in the knowledge proliferation database for the document, or Applying the fifth knowledge proliferation rule or the sixth knowledge proliferation rule to obtain two or more sets of company name, product name, or product unique name. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  (3) is a seventh knowledge proliferation rule in which a document including the unique name of the product stored in the unique name database 103 is acquired from the Web server device and stored in the knowledge proliferation database for the document. Or applying the eighth knowledge proliferation rule or the ninth knowledge proliferation rule to obtain two or more sets of company names, product names, or unique names of products. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  In (4), a document including the company name of the company database 101 and the product name stored in the product database 102 is acquired from the Web server device, and the document is stored in the knowledge proliferation database. The tenth knowledge proliferation rule is applied to obtain a set of company name, product name, and product unique name. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  (5) acquires a document including the company name in the company database 101 and the unique name of the product stored in the unique name database 103 from the Web server device, and stores the document in the knowledge proliferation database. Applying the accumulated eleventh knowledge multiplication rules, the company name, the product name, and the unique name of the product are acquired. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  (6) acquires a document including the product name stored in the product database 102 and the unique name of the product registered in the unique name database 103 from the Web server device, Applying the twelfth knowledge growth rule stored in the knowledge growth database, and obtaining a set of company name, product name, and product unique name. For example, the knowledge acquisition unit 115 sets the character string that applies to the part described as “% company name%” in each rule as the company name, and the character string that applies to the part described as “% product name%” as the product name. , A character string corresponding to the portion described as “% unique name%” is acquired as the unique name of the product.

  When the knowledge acquisition unit 115 accumulates the company name, the product name, and the unique name of the product in each corresponding database, the knowledge acquisition unit 115 sets a certain threshold for the number of acquisitions and the acquisition ratio of the used knowledge proliferation rules. Only information that exceeds the threshold may be registered in each database.

  The rule learning unit 116 searches for one or more Web server devices using one or more pieces of information existing in one or more databases, acquires one or more patterns in which one or more pieces of information appear, and the patterns are stored in advance. One or more patterns that appear more frequently as the predetermined condition is satisfied are acquired, and the one or more patterns are accumulated in the corresponding rule databases.

  Specifically, searching for one or more Web server devices using one or more information existing in each of one or more databases means company database 101, product database 102, unique name database 103, or company product. Searching a server device using one or more arbitrary information stored in one or more databases from the database 104, the company unique name database 105, or the product unique name database 106 as a search keyword.

  Searching for a Web server device is, for example, searching for character string information included in a Web page provided by a Web server device (not shown) on the Internet. Web search is performed using a search site, for example. Here, for example, transmitting a query including a search key to a so-called search engine provided by a search site and receiving information indicating a search result transmitted by the search engine is considered to be a web search, for example. . The search key can be transmitted by using an API provided by the search engine.

  To acquire one or more patterns in which one or more information appears is a pattern including one or more information existing in one or more databases used for the search from information acquired as a result of searching the server device. Is to get. The pattern may be, for example, a sentence pattern. If the target is a Web page, the pattern may be an HTML tag structure pattern, or the above-described rules may be combined. The acquisition of the pattern may be anything that can acquire the company name, the product name, and the unique name of the product, for example, the entire sentence may be acquired as a pattern, and the longest including information used for the search from multiple sentences The common partial sequence may be acquired.

To acquire one or more patterns that appear more frequently when a predetermined condition is satisfied, a condition to be registered in each rule database is determined in advance, and only patterns that satisfy the condition are acquired.
The predetermined condition is a condition that is registered in advance in each rule database. The condition may be any value that can evaluate the pattern. For example, the condition may be the number of pattern acquisitions, or the ratio occupied by the pattern among all the acquired patterns.

  To store one or more patterns in each corresponding rule database, a pattern that appears better enough to satisfy a predetermined condition is registered in the database corresponding to the pattern. The corresponding database is a company rule database 108 for a pattern for acquiring a company name, a product rule database 109 for a pattern for acquiring a product name, and a unique name rule database 110 for a pattern for acquiring a unique name of a product. A pattern for acquiring a pair of company name and product name, a pattern for acquiring a pair of company name and product unique name, a pattern for acquiring a pair of product name and product unique name, and a company name and product name If it is a pattern for acquiring a set of unique names of products, it is registered in the knowledge proliferation rule database 111.

  In addition, when the rule learning unit 116 acquires one or more patterns using one or more information existing in a database storing two or more types of information sets, the rule learning unit 116 sets the two or more types of information sets. One or more patterns are stored in one or more rule databases corresponding to two or more types of information sets using the corresponding rule database and a part of information of two or more types of information sets.

  The set of two or more types of information is information on a company name and a product name, or a company name and a product unique name, or a product name and a product unique name, or a company name, a product name, and a product unique name. The database storing a set of two or more types of information is the company product database 104, the company unique name database 105, the product unique name database 106, or the company product unique name database 107. The case where one or more patterns are acquired is a case where a pattern including two or more combinations among a company name, a product name, and a unique name of a product is acquired using a combination of two or more types of information. Using a part of information of two or more types of information means that each of the information is a combination of two types of information, each of which is a set of three types of information, and any two types All combinations of information sets are used. The rule learning unit 116 acquires a pattern using the decomposed information and registers the acquired pattern in a corresponding database. In addition, for example, when the combination of two or more types of information is a company name and a product name, the rule learning unit 116 not only acquires a pattern for acquiring the company name and product name, but also acquires a company name. Also, a pattern for acquiring the product name may be acquired, and the acquired pattern may be accumulated in the knowledge proliferation rule database 111, the company rule database 108, or the product rule database 109.

  FIG. 2 is a block diagram of the database construction device 2 in the present embodiment. The database construction device 2 includes a company product unique name database 107, a product similar group code correspondence database 21, a company similar group code unique name database 22, and a company similar group code unique name database construction unit 23.

  The product similar group code correspondence database 21 is a database that stores product names and similar group codes in association with each other. The product name is the name of a product or service that is sold or provided by a corporation and an individual, like the product name stored in the company product unique name database 107. The names of products and services here do not necessarily conform to the names of products and services described in the similar group code table. The similar group code is a code assigned to a product or service that is considered to be similar to each other in accordance with trademark examination standards, and is a code of 2 digits, 1 alphabet, and 2 digits. (Http://www.jpo.go.jp/shiryou/kijun/kijun2/ruiji_kijun9.htm)

  The product names and the similar group codes in the product similar group code correspondence database 21 are normally stored in a one-to-many relationship, and the same similar group codes are stored in association with a plurality of product names.

  The company similar group code unique name database construction unit 23 obtains a product name from a set of one or more company names, product names, and product unique names stored in the company product unique name database 107, and obtains the obtained product name. The similar group code of the product is acquired from the product similar group code correspondence database 21, and the combination of the company name, the acquired similar group code, and the unique name of the product is stored in the company similar group code unique name database 22.

  FIG. 3 is a block diagram of the trademark infringement detection apparatus 3 in the present embodiment. The trademark infringement detection device 3 includes a database 31, a reception unit 32, a trademark class determination unit 33, a product class determination unit 34, and an output unit 35.

  The database 31 may be the company product unique name database 107 constructed by the database construction device 1 or the company similar group code unique name database 22 constructed by the database construction device 2.

  The receiving unit 32 receives survey target trademark information having code specifying information that is information for specifying a similar group code of a trademark and a product. The investigation target trademark information is, for example, a set of a trademark character string that is a trademark character string and one or more code specifying information. The code specifying information is a character string that specifies a similar group code such as a product name or the similar group code itself. The trademark character string and the character string of the product name are composed of one or more characters. The characters constituting the trademark character string and the product name character string may be of any type, such as kanji, alphabet, katakana, and hiragana. In addition, two or more kinds of characters may be mixed in the characters constituting the trademark character string and the product name character string. Acceptance here means acceptance of information input from input devices such as a keyboard, mouse, touch panel, reception of information transmitted via a wired or wireless communication line, recording of an optical disk, magnetic disk, semiconductor memory, etc. It is a concept including reception of information read from a medium. Any means can be used for inputting the trademark information to be investigated, such as a numeric keypad, a keyboard, a mouse, or a menu screen. The receiving unit 32 can be realized by a device driver for input means such as a numeric keypad or a keyboard, control software for a menu screen, and the like.

  The trademark similarity determination unit 33 determines the similarity of a trademark using the unique names of one or more products included in the database 31 and the trademark included in the trademark information to be investigated.

  FIG. 4 is a block diagram of the trademark similarity determination unit 33 of the trademark infringement detection device 3 according to the present embodiment. The trademark similarity determination unit 33 includes a trademark character string acquisition unit 331, a designation acquisition unit 332, a designation display unit 333, a designation designation reception unit 334, a similarity search unit 335, and a trademark similarity determination result acquisition unit 336.

  The trademark character string acquisition unit 331 receives a trademark character string that is a trademark character string. The trademark character string is composed of one or more characters. The characters constituting the trademark character string may be of any type, such as kanji, alphabet, katakana and hiragana.

  The name acquisition unit 332 acquires name information, which is information on the name of the trademark character string, from the trademark character string received by the trademark character string acquisition unit 331. The name acquisition unit 332 may acquire a plurality of pieces of name information from one trademark character string. The name information acquired by the name acquisition unit 332 is information such as hiragana, katakana, or the like. Which of these names is determined by, for example, the specification of a similarity search unit 335 (described later) that uses the name information for searching.

  The name acquisition unit 332 may acquire the name information in any way. For example, the name acquisition unit 332 uses a dictionary having one or more character strings and information on reading of the character strings in association with each other, and uses a trademark character string stored in a storage medium (not shown). One or more pieces of name information corresponding to the trademark character string received by the acquisition unit 331 may be automatically acquired. The name acquisition unit 332 may acquire (accept) one or more pieces of name information corresponding to a trademark character string input from a user or the like via an input device (not shown). In this embodiment, a case where name information is automatically acquired will be described as an example.

  For example, the name acquisition unit 332 performs morphological analysis on the trademark character string, and acquires the reading information of the trademark character string obtained by the morphological analysis as name information. For example, the name acquisition unit 332 uses a morphological analysis to decompose a trademark character string into morphemes, acquires readings of each decomposed character string, and combines the character strings to obtain name information of the trademark character string. It is possible to obtain. In morpheme analysis, morpheme decomposition and acquisition of readings from character strings decomposed into morphemes are performed using, for example, a dictionary that associates character strings that can be morphemes with information about the readings, as described above. Done. As the morphological analysis, for example, a morphological analysis system such as “Mecab” (http://mecab.sourceforge.net/) or “ChaSen” (http://chasen.naist.jp) Etc. are available.

  The name acquisition unit 332 may acquire name information by converting a kanji, a number, a character such as an alphabet, or a word constituting a trademark character string into a reading. The process of converting kanji, numeric characters, and words into readings can be realized by, for example, a so-called re-conversion technique of a Japanese input method or a morphological analysis system technique as described above. Further, the process of converting alphabetic characters and words into readings can be realized, for example, by a technique for converting the input of the alphabet string of the Japanese input method into Japanese. A technique for converting alphabet strings into Japanese readings is disclosed in, for example, Japanese Patent Application Laid-Open No. 2009-199434. In addition, since an alphabet, a kanji, and a numerical name (reading) of a number are not necessarily one way, the name acquisition part 332 may acquire several information of various combinations. For example, the designation acquisition unit 332 may acquire a plurality of designations “IT”, “IT”, “IT”, and “It” from “IT”. Further, the designation acquisition unit 332 may acquire a plurality of designations “Yamada”, “Yamata”, “Yamaden”, “Sanden”, “Sanda”, and “Santa” from “Yamada”.

  Further, the name acquisition unit 332 accumulates in advance a character string (for example, kanji, alphabet, numbers, etc.) that can constitute a trademark character string and reading information (dictionary information) in a storage unit (not shown). In this way, one or more reading information may be acquired for each character constituting the trademark character string, and a combination of the acquired reading information may be acquired as name information.

  Further, for example, with respect to morpheme dictionary information prepared in advance having one or more character strings associated with katakana or hiragana indicating the reading, the name acquisition unit 332 may include other than katakana or hiragana included in the trademark character string. This can be realized by searching for a character string that matches the character string and sequentially acquiring reading (name) information associated with the detected character string from the dictionary information. In addition, when there are a plurality of readings for one character string or when a different name is obtained by dividing the number of characters used in the search, the name acquisition unit 332 combines these readings or changes the division. A plurality of designation information is acquired by appropriately combining the above. Such dictionary information may be held in advance by the name acquisition unit 332, for example. The name acquisition unit 332 may include a storage medium that holds such dictionary information.

  The name acquisition unit 332 may further acquire one or more divided name information by dividing the one or more name information acquired above. Specifically, the division name information is information of a part of the name information. The divided name information is used as substantially the same as the name information in, for example, a similarity search described later. For example, the name acquisition unit 332 divides each character string of the acquired name information at an arbitrary position and acquires character strings before or after being divided by the divided position, or both as divided name information. To do. The name acquisition unit 332 may sequentially change the position where the character string is divided, and acquire the divided name information according to each position where the character string is divided. For example, the name acquisition unit 332 may acquire all the character strings before and after the delimiter positions obtained when the character strings constituting the name information are divided between the characters as the divided name information. .

  The name display unit 333 displays one or more pieces of name information acquired by the name acquisition unit 332 on a display device such as a monitor (not shown). The name display unit 333 may display one or more pieces of divided name information acquired by the name acquisition unit 332. The name display unit 333 may be considered as including or not including a display device. The designation display unit 333 can be realized by display device driver software, or display device driver software and a display device.

  The designation designation accepting unit 334 accepts designation designation information for designating one or more pieces of designation information acquired by the designation acquisition unit 332 from a user or the like. For example, the designation designation information accepts designation designation information for designating one or more pieces of designation information displayed on the designation display unit 333. Further, the designation designation accepting unit 334 may accept designation designation information for designating one or more of the designation information and the divided designation information acquired by the designation acquisition unit 332 from a user or the like. For example, the designation designation accepting unit 334 accepts designation designation information for designating one or more of the designation information and the divided designation information displayed by the designation display unit 333. Incidentally, the nominal designation receiver 334, by providing a rule that defines how to specify in advance what nominal, may be may not be selected by the user. The name designation information input means may be anything such as a numeric keypad, keyboard, mouse, or menu screen. The designation designation accepting unit 334 can be realized by a device driver for input means such as a numeric keypad or a keyboard, control software for a menu screen, or the like.

  The similarity search unit 335 performs a similarity search using each name information acquired by the name acquisition unit 332. Alternatively, the similarity search unit 335 may perform a similar search using each piece of name information specified by the name designation information among the name information acquired by the name acquisition unit 332. Further, the similarity search unit 335 may perform a similarity search using each piece of nominal information acquired by the nominal acquisition unit 332 and each piece of divided nominal information. Alternatively, the similarity search unit 335 performs a similar search by using the name information and the divided name information specified by the name designation information among the name information and the divided name information acquired by the name acquisition unit 332. May be. The use of the designation name information and the division designation information specified by the designation designation information means, for example, that only the designated designation information or division designation information is used. The similarity search unit 335 selects the name information (or divided name information) from the name information (hereinafter referred to as registered name information) of the product unique name stored in advance in a storage unit such as a storage medium (not shown). Search similar registered name information. The registered name information is created and constructed in advance from the unique names in the company product unique name database 107 or the company similar group code unique name database 22.

  Note that the storage unit (not shown) in which the name information of the unique name of the product is stored may be included in the trademark kind rejection determination unit 33, or may be included in an external server device (not illustrated). . Also, the similarity search unit 335 transmits the name information or divided name information and the similar group code to a server device or the like that performs a similar search process for the unique name of a product (not shown), and executes the similar search, and the search result In such a case, it may be considered that the similarity search unit 335 has performed a similarity search.

  Further, the similarity search process by the similarity search unit 335 may be an acoustic determination process of name information. The similarity search process detects, for example, a factor that is judged not to be similar from the syllable of the nominal information, converts the factor into a penalty value, and uses the converted penalty value to determine whether the nominal information is similar The name information similar to the name information acquired from the trademark character string is detected from the name information of the unique name of the product. Here, for convenience of explanation, the division name information is also referred to as name information. Here, it is assumed that registered name information, which is name information of the unique name of the product, is stored in a storage unit (not shown) in association with the similar group code and the identification information of the unique name of the product.

  The result of the similarity search acquired by the similarity search unit 335 is information that can indicate, for example, whether or not at least similar registered name information is detected. For example, the result of the similarity search may be information indicating the presence or absence of similar registered name information, or may be information indicating one or more similar registered name information. Further, the result of the similarity search may be information indicating the number of similar registered name information. In addition, when the name information is stored in the above-described storage unit (not shown) in association with the unique name of the product or its identification information (for example, registration number), the similar search unit 335 displays similar registered name information. May be acquired from a storage unit (not shown) and output as a search result. Also, the similarity search unit 335 may output information including information on similar group codes associated with registered name information. Specific processing for performing a similar search for a title of a product unique name is known as a technique such as the above-mentioned Patent Literature 1, a title search service provided by the Japanese Patent Office, or a title search service for a commercial database. Therefore, detailed description is omitted here.

  The trademark similarity determination result acquisition unit 336 acquires, for example, the result of the similarity search by the similarity search unit 335, the company name and the similar group code specifying information related to the name information of the result of the similarity search. As for the result of the similarity search, the result of the similarity search is acquired for each name information for which the similarity search is performed. As for the result of the similarity search, for example, the result of the similarity search is acquired for each divided name information that has been subjected to the similarity search. Further, the trademark similarity determination result acquisition unit 336 may add, to each piece of evidence information, a unique name of a product that has been determined to have a similar name and identification information for the unique name of the product.

  The merchandise similarity determination unit 34 determines the similarity of the merchandise using one or more product names or one or more similar group codes included in the database 31 and code identification information included in the survey target trademark information. When the survey target trademark information holds the product name, the product type rejection determination unit 34 stores the company name and the product of the record having the product name related to the product name of the survey target trademark information from the company product unique name database 107. Get a unique name. The term “related to the product name of the survey target trademark information” means, for example, the case where the product name of the survey target trademark information and the product name of the company product unique name database 107 are the same or similar in meaning. is there. If the survey target trademark information has a similar group code, the product type determination unit 34 searches for records having the same group code as the similar group code of the survey target trademark information from the company similar group code unique name database 22. Get the company name and the unique name of the product.

  The output unit 35 outputs information on the similarity of the trademark and the similarity of the product based on the determination result of the trademark similarity determination unit 33 and the determination result of the product similarity determination unit 34. For example, the output unit 35 outputs information that can indicate whether or not similar information is detected. The information that can indicate whether or not similar information has been detected may be information indicating the presence or absence of similar registered name information in the case of the determination result of the trademark similarity determination unit 33, for example. Information indicating one or more similar registered name / call information may be used. Further, the information that can indicate whether or not similar information is detected may be information that indicates the number of similar registered name information. Further, the information that can indicate whether or not similar information is detected may be an arbitrary number of similar registered name information. Further, the output unit 35 may output the unique name of the product corresponding to similar registered name information and its identification information. Further, the output unit 35 may output all the output contents by adding a part or all of the company name, the product name, the similar group code, or the unique name of the product. Further, for example, in the case of the determination result of the product class determination unit 34, the information that can indicate whether or not similar information is detected may be information indicating the presence or absence of the corresponding code specifying information. It may be good or information indicating one or more corresponding code specifying information. Further, the information that can indicate whether or not similar information is detected may be information indicating the number of corresponding code specifying information. In addition, the output unit 35 may output all the output contents by adding one or both of the company name and the unique name.

  The output unit 35 may output all the judgment results of the trademark kind judgment unit 33 and all judgment results of the product kind judgment unit 34, or may be either one. In addition, the output unit 35 may output only the common product unique names among the results of the trademark similarity determination unit 33 and the product similarity determination unit 34.

  The output described here is a concept including display on a display, printing on paper or the like by a printer, transmission to an external device, accumulation in a recording medium, and the like. The output unit 35 may or may not include a display device. The output unit 35 can be realized by driver software of an output device or driver software of an output device and an output device.

  FIG. 5 is a flowchart showing an example of the operation of the company name acquisition unit 112 according to the present embodiment. The operation will be described below with reference to FIG.

  (Step S401) The company name acquisition unit 112 acquires a company rule from the company rule database 108. If it can be acquired, the process proceeds to step S402. If it cannot be acquired, the process ends.

  (Step S402) The company name acquisition unit 112 substitutes 1 for a counter m.

  (Step S403) The company name acquisition unit 112 determines whether or not there is an mth company rule in the company rules acquired in step S401. If there is, the process proceeds to step S404, and if not, the process ends.

  (Step S404) The company name acquisition unit 112 acquires a Web page from the Web server device. The Web page acquired at this time may be a Web page that may include a company name acquired in advance using a company rule that acquires a sentence including the company name from the company rule database 108. You may acquire a large amount at random. If the Web page can be acquired, the process proceeds to step S405. If the Web page cannot be acquired, the process proceeds to step S410.

  (Step S405) The company name acquisition unit 112 substitutes 1 for a counter n.

  (Step S406) The company name acquisition unit 112 determines whether there is an nth Web page in the Web page acquired in Step S404. If there is, the process proceeds to step S407, and if not, the process proceeds to step S410.

  (Step S407) The company name acquisition unit 112 acquires a company name for the nth Web page of the Web page acquired in Step S404 by using the mth company rule of the company rule acquired in Step S401. If it can be acquired, the process proceeds to step S408. If it cannot be acquired, the process proceeds to step S409.

  (Step S408) The company name acquisition unit 112 stores the company name acquired in step S407 in the company database 101.

  (Step S409) The company name acquisition unit 112 increments the counter n by 1. Then, the process returns to step S406.

  (Step S410) The company name acquisition unit 112 increments the counter m by 1. Then, the process returns to step S403.

  Hereinafter, a specific operation of the company name acquisition unit 112 in the present embodiment will be described. Note that the acquisition result of the company name acquisition unit 112 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The company name acquisition unit 112 acquires all rules for acquiring a company name from the company rule database 108. The company name acquisition unit 112 acquires only the company rules whose “company name” is the target when the company rule database 108 holds data related to the targets extracted by the company rules. Further, the company name acquisition unit 112 may acquire only the company rules exceeding an arbitrary threshold when the company rule database 108 holds values such as the number of acquisitions of the company rules and the acquisition ratio. In this specific example, it is assumed that five types of company rules have been acquired.

  FIG. 6 is a diagram illustrating an example of data stored in the company rule database 108. The company rule database 108 has a pattern for acquiring a company name. Here, as a specific example, company rules for acquiring a noun phrase corresponding to the part described as “% company name%” as a company name are listed.

  The company name acquisition unit 112 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The company name acquisition unit 112 acquires company names from all acquired web pages using all acquired company rules. In this case, the company name acquisition unit 112 performs the process of acquiring the company name 500 times because there are 5 types of company rules and 100 types of Web pages. Specifically, when the company rule is “<title>% company name% corporation </ title>”, the company name acquisition unit 112 displays “<title> sample trading corporation </ title>” on the acquired Web page. ”Is acquired, the“ sample trading ”part is acquired. However, the company name acquisition unit 112 acquires nothing unless the contents corresponding to the pattern are described.

  The company name acquisition unit 112 registers the company name in the company database 101 when the company name can be acquired. In this case, the company name acquisition unit 112 registers “sample trading”. Further, if the company rule database 108 holds values such as the number of acquisitions and the acquisition ratio of the company rules, the company name acquisition unit 112 may register the values together with the values in the company database 101 and acquire them. “<Title> Sample Trading Co., Ltd. </ Title>”, which is the periphery of the information, may be registered together.

  FIG. 7 is a flowchart showing an example of the operation of the product name acquisition unit 113 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S501) The product name acquisition unit 113 acquires a product rule from the product rule database 109. If it can be acquired, the process proceeds to step S502. If it cannot be acquired, the process ends.

  (Step S502) The product name acquisition unit 113 substitutes 1 for the counter m.

  (Step S503) The product name acquisition unit 113 determines whether or not the product rule acquired in step S501 includes the mth product rule. If there is, the process proceeds to step S504, and if not, the process ends.

  (Step S504) The product name acquisition unit 113 acquires a Web page from the Web server device. The Web page acquired at this time may be a Web page that may include a product name acquired in advance using a product rule that acquires a sentence including a product name from the product rule database 109. You may acquire a large amount at random. If the Web page can be acquired, the process proceeds to step S505. If the Web page cannot be acquired, the process proceeds to step S510.

  (Step S505) The product name acquisition unit 113 assigns 1 to the counter n.

  (Step S506) The product name acquisition unit 113 determines whether the Web page acquired in Step S504 includes the nth Web page. If there is, the process proceeds to step S507, and if not, the process proceeds to step S510.

  (Step S507) The product name acquisition unit 113 acquires the product name for the nth Web page of the Web page acquired in Step S504, using the mth product rule of the product rule acquired in Step S501. If it can be obtained, the process proceeds to step S508. If it cannot be obtained, the process proceeds to step S509.

  (Step S508) The product name acquisition unit 113 stores the product name acquired in step S507 in the product database 102.

  (Step S509) The product name acquisition unit 113 increments the counter n by 1. Then, the process returns to step S506.

  (Step S510) The product name acquisition unit 113 increments the counter m by 1. Then, the process returns to step S503.

  Hereinafter, a specific operation of the product name acquisition unit 113 in the present embodiment will be described. Note that the acquisition result of the product name acquisition unit 113 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The product name acquisition unit 113 acquires all rules for acquiring product names from the product rule database 109. The product name acquisition unit 113 acquires only the product rule whose “product name” is the target when the product rule database 109 holds data related to the target extracted by the product rule. In addition, the product name acquisition unit 113 may acquire only the product rules that exceed an arbitrary threshold when the product rule database 109 holds values such as the number of product rule acquisitions and the acquisition ratio. In this specific example, it is assumed that five types of product rules have been acquired.

  FIG. 8 is a diagram illustrating an example of data stored in the product rule database 109. The product rule database 109 has a pattern for acquiring product names. Here, as a specific example, product rules for acquiring a noun phrase corresponding to a portion described as “% product name%” as a product name are listed.

  The product name acquisition unit 113 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The product name acquisition unit 113 acquires product names from all acquired Web pages using all acquired product rules. In this case, the product name acquisition unit 113 performs the process of acquiring the product name 500 times because there are 5 types of product rules and 100 types of Web pages. Specifically, the product name acquisition unit 113 sets the product rule “<title> product name list </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?>% Product name% </ li>) + </ ul> ”,“ <title> product name list </ title> ... (omitted) ... <span> list </ Span> <ul> <li> <a href = ”tv. html "> TV </ li> <li> <a href =" dvd. html "> DVD player </ li> <li> <a href =" pc. If “html”> PC </ li> </ UL>... (omitted)..., “TV,” “DVD player,” “PC” is acquired. However, the product name acquisition unit 113 acquires nothing unless the content corresponding to the pattern is described.

  The product name acquisition unit 113 registers the product name in the product database 102 when the product name can be acquired. In this case, the product name acquisition unit 113 registers “TV”, “DVD player”, and “PC”. In addition, if the product rule database 109 holds values such as the number of times of product rule acquisition and the acquisition ratio, the product name acquisition unit 113 may register and acquire the values in the product database 102 together with the values. “<Title> Product name list </ title>... (Omitted)... <Span> list </ span> <ul> <li> <a href =” tv. html "> TV </ li> <li> <a href =" dvd. html "> DVD player </ li> <li> <a href =" pc. html "> PC </ li> </ ul> ... (omitted) ..." may be registered together.

  As described above, according to the present embodiment, a product name can be acquired from an arbitrary Web page.

  FIG. 9 is a flowchart showing an example of the operation of the unique name acquisition unit 114 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 601) The unique name acquisition unit 114 acquires a unique name rule from the unique name rule database 110. If it can be acquired, the process proceeds to step S602. If it cannot be acquired, the process ends.

  (Step S602) The unique name acquisition unit 114 substitutes 1 for the counter m.

  (Step S603) The unique name acquisition unit 114 determines whether or not there is an mth unique name rule in the unique name rule acquired in step S601. If there is, the process proceeds to step S604, and if not, the process ends.

  (Step S604) The unique name acquisition unit 114 acquires a Web page from the Web server device. The Web page acquired at this time is a Web page that may include the unique name of the product acquired in advance using the unique name rule for acquiring the sentence including the unique name of the product from the unique name rule database 110. There may be, and you may acquire in large quantities at random. If the Web page can be acquired, the process proceeds to step S605. If the Web page cannot be acquired, the process proceeds to step S610.

  (Step S605) The unique name acquisition unit 114 substitutes 1 for the counter n.

  (Step S606) The unique name acquisition unit 114 determines whether or not the Web page acquired in Step S604 includes the nth Web page. If there is, the process proceeds to step S607, and if not, the process proceeds to step S610.

  (Step S607) The unique name acquisition unit 114 uses the mth unique name rule of the unique name rule acquired in step S601 for the nth Web page of the Web page acquired in step S604. get. If it can be acquired, the process proceeds to step S608. If it cannot be acquired, the process proceeds to step S609.

  (Step S608) The unique name acquisition unit 114 stores the unique name of the product acquired in step S607 in the unique name database 103.

  (Step S609) The unique name acquisition unit 114 increments the counter n by 1. Then, the process returns to step S606.

  (Step S610) The unique name acquisition unit 114 increments the counter m by 1. Then, the process returns to step S603.

  Hereinafter, a specific operation of the unique name acquisition unit 114 in the present embodiment will be described. Note that the acquisition result of the unique name acquisition unit 114 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The unique name acquisition unit 114 acquires all the rules for acquiring the unique name of the product from the unique name rule database 110. When the unique name rule database 110 holds data related to the object extracted by the unique name rule, the unique name acquisition unit 114 acquires only the unique name rule whose target is “product unique name”. In addition, the unique name acquisition unit 114 may acquire only unique name rules exceeding an arbitrary threshold when the unique name rule database 110 holds values such as the number of acquisitions of unique name rules and the acquisition ratio. good. In this specific example, it is assumed that five types of unique name rules have been acquired.

  FIG. 10 is a diagram illustrating an example of data stored in the unique name rule database 110. The unique name rule database 110 has a pattern for acquiring a unique name of a product. Here, as a specific example, specific name rules for acquiring a noun phrase corresponding to a portion described as “% proper name%” as a specific name of a product are listed.

  The unique name acquisition unit 114 acquires a Web page. In this specific example, it is assumed that 100 types of Web pages are acquired at random.

  The unique name acquisition unit 114 acquires the unique name of the product from all the acquired web pages using all the acquired unique name rules. In this case, the unique name acquisition unit 114 performs processing for acquiring the unique name of the product 500 times because there are five types of unique name rules and 100 types of Web pages. Specifically, the unique name acquisition unit 114 has a unique name rule of “<title> product list </ title>. *? <Span. *?> List </ span> <ul. *?> (<Li. *?> <A. *?>% Unique name% </a> </ li>) + </ ul> ”,“ <title> product list </ title>... ( Omitted) ... <span> list </ span> <ul> <li> <a href = ”sc. html "> sample camera </ li> <li> <a href =" sc2. If “html”> sample camera 2 </ li> </ ul>... (omitted)..., “sample camera” and “sample camera 2” are acquired. However, the unique name acquisition unit 114 acquires nothing unless the content corresponding to the pattern is described.

  When the unique name of the product can be acquired, the unique name acquisition unit 114 registers it in the unique name database. In this case, the unique name acquisition unit 114 registers “sample camera” and “sample camera 2”. If the unique name rule database 110 holds values such as the number of acquisitions of the unique name rule and the acquisition ratio, the unique name acquisition unit 114 may register the unique name rule together with the values in the proper name database 103. “<Title> product list </ title>... (Omitted)... <Span> list </ span> <ul> <li> <a href =” sc. html "> sample camera </ li> <li> <a href =" sc2. html "> sample camera 2 </ li> </ ul> ... (omitted) ..." may be registered together.

  As described above, according to the present embodiment, a unique name of a product can be acquired from an arbitrary Web page.

  FIG. 11 is a flowchart showing an example of the operation of the knowledge acquisition unit 115 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 701) The knowledge acquisition unit 115 acquires a knowledge proliferation rule from the knowledge proliferation rule database 111. If it can be acquired, the process proceeds to step S702. If it cannot be acquired, the process ends.

  (Step S702) The knowledge acquisition unit 115 substitutes 1 for a counter m.

  (Step S703) The knowledge acquisition unit 115 determines whether or not the knowledge multiplication rule acquired in step S701 includes the mth knowledge multiplication rule. If there is, the process proceeds to step S704, and if not, the process ends.

  (Step S704) The knowledge acquisition unit 115 acquires all information used for the m-th knowledge proliferation rule of the knowledge proliferation rule acquired in step S701 from each database. Specifically, the information used for the knowledge proliferation rule is obtained by acquiring the company name from the company database 101 in the case of the first to third knowledge proliferation rules, and in the case of the fourth to sixth knowledge proliferation rules. The product name is acquired from the product database 102, the unique name of the product is acquired from the unique name database 103 in the case of the seventh to ninth knowledge proliferation rules, and the company name and the company name are acquired from the company product database 104 in the case of the tenth knowledge proliferation rule. The product name is acquired. In the case of the eleventh knowledge proliferation rule, the company name and the product unique name are acquired from the company unique name database 105. In the case of the twelfth knowledge proliferation rule, the product name and the product name are acquired from the product unique name database 106. Get the unique name of the product.

  (Step S705) The knowledge acquisition unit 115 substitutes 1 for a counter n.

  (Step S706) The knowledge acquisition unit 115 determines whether there is n-th information in the information used for the knowledge proliferation rule acquired in step S704. If there is, the process proceeds to step S707, and if not, the process proceeds to step S713.

  (Step S707) The knowledge acquisition unit 115 acquires a Web page from the Web server device. The Web page acquired at this time acquires a Web page including the nth information among the information used for the knowledge proliferation rule acquired in step S704. If the Web page can be acquired, the process proceeds to step S708. If the Web page cannot be acquired, the process proceeds to step S714.

  (Step S708) The knowledge acquisition unit 115 assigns 1 to the counter o.

  (Step S709) The knowledge acquisition unit 115 determines whether or not the web page acquired in step S707 includes the o-th web page. If there is, the process proceeds to step S710, and if not, the process proceeds to step S714.

  (Step S710) The knowledge acquisition unit 115 uses the m-th knowledge proliferation rule of the knowledge proliferation rule acquired in Step S701 for the o-th Web page of the Web page acquired in Step S707. , Or get the unique name of the product. If it can be acquired, the process proceeds to step S711. If it cannot be acquired, the process proceeds to step S712.

  (Step S711) The knowledge acquisition unit 115 accumulates the company name, the product name, or the unique name of the product acquired in Step S710 in the corresponding database.

  (Step S712) The knowledge acquisition unit 115 increments the counter o by 1. Then, the process returns to step S709.

  (Step S713) The knowledge acquisition unit 115 increments the counter m by 1. Then, the process returns to step S703.

  (Step S714) The knowledge acquisition unit 115 increments the counter n by 1. Then, the process returns to step S706.

  Hereinafter, a specific operation of the knowledge acquisition unit 115 in the present embodiment will be described. Note that the acquisition result of the knowledge acquisition unit 115 shown in this specific example is prepared for convenience of explanation, and does not indicate actual data.

  The knowledge acquisition unit 115 acquires all rules for acquiring a company name, a product name, or a unique name of a product from the knowledge multiplication rule database 111. The knowledge acquisition unit 115 acquires only the knowledge proliferation rule whose “name” is the object when the knowledge proliferation rule database 111 holds data related to the object extracted by the knowledge proliferation rule. Note that the knowledge acquisition unit 115 may acquire only knowledge proliferation rules that exceed an arbitrary threshold value when values such as the number of acquisitions and the acquisition ratio of knowledge proliferation rules are held. In this specific example, it is assumed that five types of knowledge proliferation rules have been acquired.

  FIG. 12 is a diagram showing an example of data stored in the knowledge growth rule database 111. The knowledge proliferation rule database 111 has a pattern for acquiring a company name, a product name, or a unique name of a product. The knowledge proliferation rule in this specific example is that the noun phrase corresponding to the part described as “% company name%” is the noun phrase corresponding to the part described as “% product name%”. This is a knowledge multiplication rule for acquiring a noun phrase corresponding to the part described as the product name “% proper name%” as the product proper name. In this specific example, instead of creating a database for each type of knowledge proliferation rule, the knowledge proliferation rule type is held in the knowledge proliferation database.

  The knowledge acquisition unit 115 acquires information to be used by the acquired knowledge proliferation rule. For example, when acquiring the first knowledge proliferation rule, the knowledge acquisition unit 115 acquires all the company names described in the company database 101. Note that the knowledge acquisition unit 115 may not acquire a company name equal to or less than an arbitrary threshold when the company database 101 holds values such as the number of acquisitions and the acquisition ratio of the company name. In this specific example, it is assumed that ten types of company names have been acquired.

  The knowledge acquisition unit 115 acquires a web page. For example, in this specific example, it is assumed that the knowledge acquisition unit 115 acquires 100 types of Web pages for each acquired company name.

The knowledge acquisition unit 115 acquires company names from all acquired Web pages using all acquired knowledge proliferation rules. In this case, the knowledge acquisition unit 115 has 5 types of knowledge proliferation rules, 10 types of information used by the knowledge proliferation rules, and 100 types of Web pages, so 1 of 5000 unique company names, product names, and product unique names. Or two types of information are acquired.
Specifically, the knowledge acquisition unit 115 sets the knowledge multiplication rule to “<title>% company name% (handling)? Merchandise </ title>. *? <Span. *?> List </ span><ul.*?>(<Li.*?><A.*?>% Product name% </a></li>) + </ ul> ”is the first knowledge proliferation rule, and the information used is“ sample In the case of “Industry”, “<title> Sample Industry Products </ title> ... (omitted) ... <span> list </ span><ul><li><a href = "Scr. html ”> screw </a></li><li><a href =“ driv. If “html”> driver </a></li></ul>... (omitted)..., “screw” and “driver” are acquired. The knowledge acquisition unit 115 acquires nothing unless the content corresponding to the pattern is described.

  When new information can be acquired, the knowledge acquisition unit 115 registers it in each corresponding database. In this case, the knowledge acquisition unit 115 registers “screw” and “driver” in the product database 102, “sample industry” and “screw”, and “sample industry” and “driver” in the company product database 104. The knowledge acquisition unit 115 may register the knowledge proliferation rule database 111 together with the values when the knowledge proliferation rule database 111 holds values such as the number of acquisitions and the acquisition ratio of the knowledge proliferation rules. There are certain “<title> sample industry products </ title>... (Omitted)... <Span> list </ span> <ul> <li> <a href =” scr. html ”> screw </a> </ li> <li> <a href =“ driv. html ”> driver </a> </ li> </ ul>... (omitted)... Note that the knowledge acquisition unit 115 performs the above processing on all acquired knowledge proliferation rules.

  As described above, according to the present embodiment, a company name, a product name, or a unique name of a product can be acquired from an arbitrary Web page.

  13 and 14 are flowcharts illustrating an example of the operation of the rule learning unit 116 according to the present embodiment. The operation will be described below with reference to FIGS.

  (Step S801) The rule learning unit 116 substitutes 0 for the counter m.

  (Step S802) The rule learning unit 116 increments the counter m by 1.

  (Step S803) The rule learning unit 116 branches the process depending on the value of the counter m. When the value of m is 1, the process proceeds to step S804. When the value of m is 2, the process proceeds to step S805. When the value of m is 3, the process proceeds to step S806. When the value of m is 4, the process proceeds to step S807. Proceed to step S810 when the value of m is 5, proceed to step S813 when the value of m is 6, proceed to step S816 when the value of m is 7, and end when the value of m is 8 To do.

  (Step S804) The rule learning unit 116 substitutes “company database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S805) The rule learning unit 116 substitutes “product database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S806) The rule learning unit 116 substitutes “unique name database” for the variable o and “unique name of the product” for the variable d. Then, the process proceeds to step S823.

  (Step S807) The rule learning unit 116 substitutes “company product database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S808) The rule learning unit 116 substitutes “company product database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S809) The rule learning unit 116 substitutes “company product database” for the variable o and “company name / product name” for the variable d. Then, the process proceeds to step S823.

  (Step S810) The rule learning unit 116 substitutes “company unique name database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S811) The rule learning unit 116 substitutes “company unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S812) The rule learning unit 116 substitutes “company unique name database” for the variable o and “company name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S813) The rule learning unit 116 substitutes “product unique name database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S814) The rule learning unit 116 substitutes “product unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S815) The rule learning unit 116 substitutes “product unique name database” for the variable o and “product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S816) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name” for the variable d. Then, the process proceeds to step S823.

  (Step S817) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “product name” for the variable d. Then, the process proceeds to step S823.

  (Step S818) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S819) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / product name” for the variable d. Then, the process proceeds to step S823.

  (Step S820) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S821) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S822) The rule learning unit 116 substitutes “company product unique name database” for the variable o and “company name / product name / unique name” for the variable d. Then, the process proceeds to step S823.

  (Step S823) The rule learning unit 116 acquires information related to information stored in the variable d from the database stored in the variable o. For example, when “company database” is stored in the variable o and “company name” is stored in the variable d, the rule learning unit 116 acquires all the company names from the company database 101. In addition, the rule learning unit 116, when the information stored in the variable d is registered with two types of information separated by “·”, such as “company name / product name”, before and after “·” Split the column and get both front and back information pairs. In the case of “company name / product name”, the rule learning unit 116 acquires a set of “company name” and “product name”. If it can be acquired, the process proceeds to step S824, and if it cannot be acquired, the process proceeds to 829.

  (Step S824) The rule learning unit 116 substitutes 1 for the counter h.

  (Step S825) The rule learning unit 116 determines whether or not there is h-th information in the information regarding d acquired from o acquired in step S823. If there is, the process proceeds to step S826, and if not, the process proceeds to step S830.

  (Step S826) The rule learning unit 116 acquires a Web page from the Web server device. The Web page acquired at this time acquires a Web page including the h-th information regarding d acquired from o acquired in step S823. If the Web page can be acquired, the process proceeds to step S827. If the Web page cannot be acquired, the process proceeds to step S829.

  (Step S827) The rule learning unit 116 replaces the character string of the h-th information related to d acquired from o acquired in Step S823 from the Web page acquired in Step S826 with a character string that understands the content of the information. For example, the rule learning unit 116 replaces the company name portion with “% company name%”.

  (Step S828) The rule learning unit 116 temporarily stores the Web page replaced in Step S827 in a storage medium such as a memory (not shown).

  (Step S829) The rule learning unit 116 increments the counter h by 1.

  (Step S830) The rule learning unit 116 acquires patterns from all the Web pages stored in Step S828. If the pattern can be acquired, the process proceeds to step S831, and if the pattern cannot be acquired, the process proceeds to step S833.

  (Step S831) The rule learning unit 116 selects and acquires patterns extracted five or more times from the patterns acquired in Step S830. If the pattern can be acquired, the process proceeds to step S832, and if the pattern cannot be acquired, the process proceeds to step S833.

  (Step S832) The rule learning unit 116 stores the pattern acquired in step S831 in the rule database corresponding to the information stored in the variable d. Specifically, the rule learning unit 116 is “company rule database” if the value of “d” is “company name”, “product rule database” if “d” is “product name”, and “unique” if it is “unique name”. If “name rule database”, “company name / product name”, “company name / unique name”, “product name / unique name” are stored in the “knowledge proliferation rule database”.

  (Step S833) The rule learning unit 116 proceeds to a process next to the process executed before proceeding to step S823.

  Hereinafter, a specific operation of the rule learning unit 116 in the present embodiment will be described. Note that the acquisition result of the rule learning unit 116 shown in this specific example is prepared for convenience of explanation, and does not indicate actual data.

  The rule learning unit 116 selects the company name or the product name from the company database 101, the product database 102, the unique name database 103, the company product database 104, the company unique name database 105, the product unique name database 106, and the company product unique name database 107. Or a unique name of a product or a combination thereof is acquired as much as possible, and a pattern for extracting information acquired from each database is acquired from a Web page on which the acquired information is described.

  As a specific example, a case where the rule learning unit 116 acquires a company name from the company database 101 and learns a rule stored in the company rule database 108 will be described. When learning rules other than those described above, the rules shall be read as appropriate.

  First, the rule learning unit 116 acquires all company names from the company database 101. Note that the rule learning unit 116 may not acquire a company name equal to or less than an arbitrary threshold when the company database 101 holds values such as the number of acquisitions and the acquisition ratio of the company name. In this specific example, it is assumed that the rule learning unit 116 has acquired 10 types of company names.

  The rule learning unit 116 acquires a web page. For example, in this specific example, assume that 100 types of Web pages are acquired for each acquired company name.

  The rule learning unit 116 replaces the company name used for acquisition with “% company name%” from the acquired Web page. In this case, for example, the rule learning unit 116 uses “sample trading” to read “<table> <tbody> <tr> <td> company name </ td> <td> sample trading corporation </ td> </ tr > <Tr> <td> Capital </ td> <td> 10,000,000 Yen </ td> </ tr> <text> <table> <tbody> <tr> <td > Company name </ td> <td>% Company name% Inc. </ td> </ tr> <tr> <td> Capital </ td> <td> 10,000,000 Yen </ td> <// tr> ”.

  The rule learning unit 116 acquires a pattern from the acquired web page. The rule learning unit 116 extracts two or more Web pages from the acquired Web pages, and acquires the longest common subsequence including “% company name%” used for acquisition. Acquisition of the longest common portion is performed by all combinations of selecting two from the acquired Web pages. The rule learning unit 116, for example, “<table> <tbody> <tr> <td> company name </ td> <td>% company name% corporation </ td> </ tr> <tr> <td> capital </ Td> <td> 10,000,000 Yen </ td> </ tr> ”and“ <table> <tr> <td> Company name </ td> <td>% Company name% Inc. </ Td ” > </ Tr> <tr> <td> founded </ td> <td> January 1, 1950 </ td> </ tr> ” “<Tr> <td> company name </ td> <td>% company name% Inc. </ Td> </ tr> <tr> <td> which is the longest character string holding“% company name% ” "As a pattern. In this case, the rule learning unit 116 acquires the pattern 49500 times because there are 10 types of company names and 100 types of Web pages for the company name. Note that the rule learning unit 116 may acquire a pattern from a combination of two or more Web pages, and the pattern acquisition target may be a sentence that does not include an HTML tag, or may be a method other than the longest common subsequence. May be.

  The rule learning unit 116 registers only frequently used patterns from the acquired patterns in the knowledge multiplication database. Further, the rule learning unit 116 registers a pattern that appears at a certain frequency, for example, five times or more, in the knowledge proliferation rule database 111.

  As described above, according to the present embodiment, it is possible to acquire a pattern for acquiring a company name, a product name, or a unique name of a product from an arbitrary Web page.

  FIG. 15 is a flowchart showing an example of the operation of the company similar group code unique name database construction unit 23 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S <b> 901) The company similar group code unique name database construction unit 23 acquires information on the company name, product name, and product unique name from the company product unique name database 107. If it can be acquired, the process proceeds to step S902. If it cannot be acquired, the process ends.

  (Step S902) The company similar group code unique name database construction unit 23 substitutes 1 for a counter m.

  (Step S903) The company similar group code unique name database construction unit 23 determines whether or not the information acquired in step S901 includes the m-th information. If there is, the process proceeds to step S904, and if not, the process ends.

  (Step S904) The company similar group code unique name database construction unit 23 acquires a similar group code corresponding to the product name from the product similar group code correspondence database 21 using the product name of the m-th information acquired in step S901 as a key. To do.

  (Step S905) The company similar group code unique name database construction unit 23 uses the m-th information company name acquired in step S901, the unique name of the product, and the similar group code acquired in step S904 as one record. Register in the group code unique name database 22.

  (Step S906) The company similar group code unique name database construction unit 23 increments the counter m by 1.

  Hereinafter, a specific operation of the company similar group code unique name database construction unit 23 in the present embodiment will be described. Note that the acquisition result of the company similar group code unique name database construction unit 23 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data.

  The company similar group code unique name database construction unit 23 acquires a similar group code from the similar group code correspondence database using the product name of the company product unique name database 107 as a key, and constructs the company similar group code unique name database 22. .

  First, the company similar group code unique name database construction unit 23 acquires all the company name, product name, and product unique name from the company product unique name database 107. In addition, the company similar group code unique name database construction unit 23 does not acquire a company name below an arbitrary threshold when the company product unique name database 107 holds values such as the number of acquisitions and the acquisition ratio. Also good. In this specific example, it is assumed that ten types of combinations of company name, product name, and product unique name have been acquired.

  Next, the company similar group code unique name database construction unit 23 acquires a similar group code corresponding to each product name from the product similar group code correspondence database 21. Specifically, the company similar group code unique name database construction unit 23 acquires not only a perfect match of product names but also a similar group code that conceptually matches product names such as “TV” and “TV”, for example. You may do it.

  Then, the company similar group code unique name database construction unit 23 registers the set of the acquired company name, similar group code, and product unique name in the company similar group code unique name database 22.

  As described above, according to the present embodiment, the company similar group code unique name that holds the company name, the similar group code, and the unique name of the product using the product name as a key from the company product unique name database 107 and the similar group code correspondence database. A database 22 can be constructed.

  FIG. 16 is a flowchart showing an example of the operation of the trademark infringement detection apparatus 3 according to the present embodiment. Hereinafter, the operation will be described with reference to FIG.

  (Step S1001) The receiving unit 32 determines whether or not the investigation target trademark information, specifically, the trademark character string and the product name or the similar group code has been received. If accepted, the process proceeds to step S1002, and if not accepted, the process returns to step S1001.

  (Step S1002) The trademark character string acquisition unit 331 acquires a trademark character string from the investigation target trademark information. If it can be acquired, the process proceeds to step S1003. If it cannot be acquired, the process proceeds to step S1017.

  (Step S1003) The name acquisition unit 332 acquires one or more pieces of name information using the trademark character string acquired in step S1002. For example, the name acquisition unit 332 performs morphological analysis on the trademark character string, reads each character constituting the trademark character string from a dictionary or the like in which a character reading prepared in advance is stored, and combines these combinations. By acquiring, name information is acquired.

  (Step S1004) The name acquisition unit 332 substitutes 1 for the counter m.

  (Step S1005) The name acquisition unit 332 determines whether there is m-th name information in the name information acquired in step S1003. If there is, the process proceeds to step S1006, and if not, the process proceeds to step S1011.

  (Step S1006) The name acquisition unit 332 substitutes 1 for the counter n.

  (Step S1007) The name acquisition unit 332 determines whether or not the m-th name information includes the n-th delimiter position. For example, when setting between all characters of the nominal information as the separation position, the nth separation position is, for example, between the nth character and the (n + 1) th character from the beginning of the nominal information. is there. For this reason, the name acquisition unit 332 may determine whether, for example, the mth name information includes the (n + 1) th character from the top. If there is an nth delimiter position, the process proceeds to step S1008, and if not, the process proceeds to step S1010. The name acquisition unit 332 sets the name information by setting the interval between reading of each character constituting the trademark character string or reading of the morpheme constituting the trademark character string in the name information. You may make it divide | segment. For example, when the nominal information “ABC” is acquired by combining the readings “A”, “B”, and “SE” of each character of the trademark character string “ABC”, the nominal acquisition unit 332 is a position that can be set as a separation position. May be between “A” and “B” and between “B” and “SE”.

  (Step S1008) The name acquisition unit 332 divides the mth name information at the nth position, and acquires divided name information that is a character string before and after the mth name information. It should be noted that a character string preceded by a predesignated character that cannot be a trademark name is preferably not acquired as divided name information. The character designated in advance is, for example, a long sound or “n”. The acquired divided name information is temporarily stored in a storage medium such as a memory (not shown). If other divided name information is already stored, it is added.

  (Step S1009) The name acquisition unit 332 increments the counter n by 1. Then, the process returns to step S1007.

  (Step S1010) The name acquisition unit 332 increments the counter m by 1. Then, the process returns to step S1005.

  (Step S1011) The name display unit 333 displays the name information acquired in step S1003 and the divided name information acquired in step S1008. Note that the designation display unit 333 may skip the display of this step and the reception of step S1012 when selecting designation of designation by a specific rule without causing the user to select.

  (Step S1012) The designation designation receiving unit 334 determines whether designation designation information for designating at least one of designation designation information or divided designation designation information has been accepted. If accepted, the designation display unit 333 temporarily stores the designation information and designation call information specified by the designation designation information in a storage medium such as a memory, and proceeds to step S212. If not accepted, the designation display unit 333 proceeds to step S1012. Return to.

  (Step S1013) The similarity search unit 335 substitutes 1 for the counter k.

  (Step S1014) The similarity search unit 335 determines whether there is k-th name information or divided name information in the combination of the name information and the divided name information specified by the name designation information received in Step S1012. to decide. If there is, the process proceeds to step S1015, and if not, the process proceeds to step S1017.

  (Step S1015) The similarity search unit 335 searches the database 31 using the k-th name information or divided name information, and performs a similarity search. Then, the similarity search unit 335 accumulates the results of the similarity search in a storage medium (not shown). As a result of the similarity search, for example, k-th name information or divided name information used for the similarity search, a name determined to be similar, a company name associated with the name, and a unique name of the product It is information to include.

  (Step S1016) The similarity search unit 335 increments the counter k by 1. Then, the process returns to step S1014.

  (Step S1017) The merchandise type determination unit 34 selects a database to be used based on the survey target trademark information acquired in step S1001 from the database 31. Specifically, the merchandise inequality determination unit 34 selects the company similar group code unique name database 22 when the survey target trademark information includes the similar group code, and when the merchandise name is included. The company product unique name database 107 is selected. If it can be selected, the process proceeds to step S1018. If it cannot be selected, the process proceeds to step S1019.

  (Step S1018) The product type determination unit 34 determines the product type using the product name or similar group code of the investigation target trademark information acquired from the database 31 of step S1017 in step S1001 as a key. Then, the merchandise similarity determination unit 34 accumulates the result of similarity determination in a storage medium (not shown). The result of similarity determination is, for example, the product name or similar group code used for similarity determination, the product name or similar group code determined to match, and the company name and product associated with the designation. Information including the unique name.

  (Step S1019) The trademark infringement detection device 3 acquires information having a common product unique name from the similar search result accumulated in step S1015 and the information accumulated in step S1018.

  (Step S1020) The output unit 35 outputs the information acquired in step S1019. Then, the process ends.

  Hereinafter, a specific operation of the trademark infringement detection device 3 in the present embodiment will be described. The data shown in this specific example is prepared for the sake of explanation.

  First, the user wants to judge the possibility of infringement of “wire noodles” which are noodles of Chinese noodles, and therefore, the trademark character string “wire noodles” and a similar group code “32F03” of Chinese noodles of noodles. ”Is input to the trademark infringement detection device 3 by operating a keyboard or the like. The receiving unit 32 receives the trademark character string “wire noodle”. Then, the trademark character string acquisition unit 331 acquires a trademark character string, and the product category determination unit 34 acquires a similar group code.

  The name acquisition unit 332 acquires the name information of the trademark character string “wire noodle” received by the trademark character string acquisition unit 331. Here, first, morphological analysis is performed on the trademark character string to obtain reading information for each morpheme. For example, “wire noodle” is divided into two morphemes “wire” and “noodle” by morpheme analysis. “Harigane” is read from the morpheme “wire”, and “noodle” is read from the morpheme “noodle”. “Men” is acquired. Since only one reading is acquired for each morpheme, “harnessing” obtained by simply combining these readings is acquired as one of the designation information of “wire noodles”.

  Furthermore, the name acquisition unit 332 acquires a reading of each character of the trademark character string “wire noodle” from a dictionary (not shown) in which kanji readings prepared in advance are stored. For example, it is assumed that readings “harness” and “shin” are acquired from “needle”. In addition, it is assumed that “Kin”, “Kogane”, “Kon”, and “Kane” are acquired from “Gold”. It is also assumed that “Men” is acquired from “Noodles”. Then, the name acquisition unit 332 acquires the name information by combining the acquired reading of the characters along the sequence of the characters as the acquisition source. For example, the designation acquisition unit 332 acquires designation information such as “Harikinmen”, “Harikomenmen”, “Harikonmen”, “Harikanmen”, “Shinkinmen”, “Cincoganmen”, and the like. Note that information that matches the name information acquired by morphological analysis is deleted.

  Next, the name acquisition unit 332 acquires the divided name information by dividing the name information acquired above. For example, with respect to the designation information “Hariganemen”, first, by dividing the designation information by setting the division position between “Ha” of the first character and “Li” of the second character, the divided designation of “Ha” is given. Information and division name information “Rigane men” are acquired. Further, the name acquisition unit 332 divides the name information by setting the division position between “Li” of the second character and “Ga” of the third character, thereby dividing the divided name information of “Hari” and “ Divided name information “Ganemen” is acquired. Similarly, the name acquisition unit 332 acquires “hariga”, “nemen”, “harness”, “men”, and “harness”. However, “n” is not acquired as divided name information. Similarly, the name acquisition unit 332 acquires divided name information for other name information. Note that the name acquisition unit 332 deletes the acquired divided name information except for one duplicated name information.

  Then, the name display unit 333 outputs the acquired name information and divided name information to a monitor or the like.

  FIG. 17 is a diagram showing a display example of the name information and the divided name information by the name display unit 333. In this case, the name information and the divided name information are used without distinction in the later-described similar search, and therefore, the name information and the divided name information are displayed as names without distinction.

  Next, it is assumed that the user checks the check box next to the name information and the divided name information shown in FIG. 17 that he / she wants to use for similarity search and presses the “designation complete” button 31. . For example, if the check boxes next to “Harigamimen”, “Harigami”, and “Harikinmen” are checked as shown in FIG. 17 and the “Designation Complete” button 31 is pressed, the designation designation receiving unit 334 is checked. “Hariganemen” and “Harikinmen”, which are the name information, and “Harigane”, which is the divided name information which is checked, are received.

  Next, the similarity search unit 335 searches the database 31 using the name information and the divided name information received by the name designation reception unit 334, and performs a name similarity search. Since the similarity search for names is a known technique, a detailed specific example is omitted. For example, as a result of the similarity search, for the name information “Hariganmen” and the divided name information “Harigane”, a unique name of a product with a similar name was not detected, but for the name information “Harikinmen”, a product with a similar name Suppose that a unique name is detected. Specifically, it is assumed that a unique name of a product whose name is “Hurricington” is detected. The trademark similarity determination result acquisition unit 336 temporarily stores the detection result in a storage medium (not shown).

  Then, the trademark similarity determination result acquisition unit 336 acquires the result of the similarity search.

  Next, the merchandise similarity determination unit 34 acquires information from the company similar group code unique name database 22 using the similar group code acquired from the reception unit 32 that determines the similarity of the product. In the case of “32F03”, the product type determination unit 34 acquires all the information related to the processed grain product regardless of the “wire surface”, but acquires all.

  And the output part 35 acquires the information with the common name of goods among the judgment result of the trademark kind judgment part 33, and the judgment result of the goods kind judgment part 34, The company name and the unique name of goods are obtained. Output. At this time, the output unit 35 may output the information that identifies the company in the company database 101, the certainty of the company name, or other information that is held.

  As described above, according to the present embodiment, the name information of the trademark character string is automatically acquired from the survey target trademark character string designated by the user, and the similarity search using this can be performed. It is possible to provide the user with information that can be used to determine the overall infringement status.

  In addition, according to the present embodiment, the divided name information obtained by dividing the name information is acquired and the similarity search using the divided name information can be performed. For example, when the trademark input by the user is a combined trademark. Can automatically acquire a name when the combined trademark is divided, and an appropriate similarity search using the divided name information can be performed. In addition, it is possible to easily input the divided name information and to prompt the user to perform a similar search using the divided name information. In addition, by acquiring the division name information, an appropriate similarity search using only the main part of the trademark becomes possible. Further, it is possible to easily input the divided name information and to prompt the user to perform a similar search using the main part.

(Embodiment 5)
The second embodiment will be described more clearly. The present embodiment is characterized in that the similarity of the product can be determined using the similar group code even when the investigation target trademark information holds the product name and does not hold the similar group code.

  FIG. 20 is a block diagram of the trademark infringement detection device 4 in the present embodiment. The trademark infringement detection apparatus 4 includes a database 401, a reception unit 32, a trademark class rejection determination unit 33, a product class determination unit 402, and an output unit 35. Since the reception unit 32, the trademark similarity determination unit 33, and the output unit 35 are the same as those in the fourth embodiment, description thereof is omitted.

  The database 401 can store one or more sets of company names, product names, and product unique names, which are information acquired from one or more Web server devices (not shown). The database 401 may be any information obtained from one or more Web server devices, may be constructed using the database construction device 1, and is searched by using a product name of random or surveyed trademark information as a keyword. Information may be acquired from a Web page collected using a specific rule. The specific rule is, for example, a rule stored in each rule database of the database construction device 1.

  The commodity similarity determination unit 402 performs commodity similarity determination using one or more commodity names included in the database 401 and the code identification information included in the investigation target trademark information, and the investigation target stored in the database 401 The company name, product name, and product unique name similar to the trademark information are acquired. The product category determination unit 402 includes a product code management unit 4021, a code identification information acquisition unit 4022, and a product category determination unit 4023. The code specifying information in the present embodiment may be a product name, a number for referring to a similar group code managed in a storage unit (not shown), or the like, and a similar group code table May be the product or service name (for example, “Chemical” in the case of 01A01), the product / service name in the similar group code table, or the similar group code itself.

  The product code management unit 4021 manages the correspondence between the product name and the code specifying information. The product name to be managed and the code specifying information may be acquired from the description of the similar group code using a specific rule, or may be acquired from a Web page acquired from a Web server device (not shown) using a specific rule. Often, a person may input manually. The specific rule is, for example, a rule for acquiring “04A01” as the code identification information and “washing powder” and “shampoo” as the product name from a sentence such as “04A01 is for example washing powder or shampoo”. From the HTML structure such as “<table> <tr> <td> dye </ td> <td> indigo / akane </ td> </ tr> </ table>” as code specifying information The rule may be to acquire “dye” and “indigo” and “Akane” as trade names. The product code management unit 4021 may be configured to acquire a similar group code by recursively searching for the correspondence. Specifically, when there are two sets of information, code specific information “01A01” and product name “Chemical”, code specific information “Chemical” and product name “Ammonia water”, the product name “Ammonia water” You may manage so that code specific information "01A01" may be acquired.

  The code specifying information acquisition unit 4022 acquires code specifying information corresponding to one or more product names included in the database 401, using the code specifying information corresponding to the product managed by the product code management unit 4021. The code specifying information to be acquired may be arbitrary code specifying information corresponding to the product name of the database 401, may be a similar group code acquired as a result of recursively acquiring the correspondence, It may be all the code specifying information acquired in the process of acquiring the correspondence recursively until it arrives. The code specifying information acquisition unit 4022 can be usually realized by an MPU, a memory, or the like. The processing procedure of the code specifying information acquisition unit 4022 is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The product type determination unit 4023 makes a product type determination using the code specification information acquired by the code specification information acquisition unit 4022 and the code specification information included in the investigation target trademark information. In determining whether a product is similar, the code specifying information included in the trademark information to be investigated and the code specifying information acquired by the code specifying information acquiring unit 4022 are determined to be the same or included. In this case, the company name, the product name, and the unique name of the product corresponding to the information are acquired. The merchandise kind determination unit 4023 can usually be realized by an MPU, a memory, or the like. The processing procedure of the merchandise kind determination unit 4023 is usually realized by software, and the software is recorded in a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  The operation of the trademark infringement detection apparatus 4 according to the present embodiment is a modification of only the process of step S1018 in FIG.

  Hereinafter, a specific operation of the merchandise inequality determination unit 402 in the trademark infringement detection device 4 according to the present embodiment will be described. It should be noted that the output result of the commodity similarity determination unit 402 shown in this specific example is prepared for the sake of explanation, and does not indicate actual data. Here, as an example, the code specifying information of the investigation target trademark information received by the receiving unit 32 is “beer”, and the database 401 has a company name “A company”, a product name “Lugger beer”, and a unique product name. It is assumed that at least one set of information whose name is “B beer” is stored, and the product code management means 4021 has the code specifying information “28A02” and the product name “beer”, the code specifying information “beer” and the product name “ It is assumed that at least two sets of information “Lager beer” are stored.

  The code specifying information acquisition unit 4022 acquires information on the product name from the database 401. For example, the code specifying information acquisition unit 4022 acquires information such as “Lugger beer”. And the code specific information acquisition means 4022 acquires code specific information from each product name using the information which the product code management means 4021 manages. Specifically, the code specifying information acquisition unit 4022 acquires all code specifying information from “Lager Beer” to “Beer” and “Beer” to “28A02” until the similar group code is acquired, and corresponds to the product name. In addition, it is stored in a storage unit (not shown).

  The product type determination unit 4023 acquires the information holding the code specifying information identical to the “beer” included in the survey target trademark information received by the receiving unit 32 by the code specifying information acquiring unit 4022, and stores it in a storage unit (not shown). Searching from the stored information, the company name and the unique name of the product stored in the database 401 are acquired. Specifically, the product type determination unit 4023 acquires “lager beer” associated with “beer” and “28A02”, the company name is “Company A”, the product name is “Lager beer”, The information that the unique name of the product is “B beer” is acquired.

  As described above, according to the present embodiment, a trademark taking account of a similar group code without specifying a similar group code for a set of a company name, a product name, and a product unique name acquired from a Web server device. Thus, it is possible to provide the user with information that can be used to determine the trademark infringement status.

  In the present embodiment, the code specifying information is acquired by the code specifying information acquiring unit 4022, but the code specifying information may be acquired in advance using a processing unit (not shown) and stored in the database 401. good.

  Further, in the fourth embodiment and the fifth embodiment, the company name acquisition unit 112, the product name acquisition unit 113, the unique name acquisition unit 114, the knowledge acquisition unit 115, the rule learning unit 116, and the company similar group code unique name database construction Unit 23, product category determination unit 34, trademark character string acquisition unit 331, name acquisition unit 332, similarity search unit 335, trademark class determination result acquisition unit 336, code identification information acquisition unit 4022, product category determination unit 4023 Usually, it can be realized from an MPU, a memory, or the like. The processing procedure of the product name acquisition unit 113 and the like is usually realized by software, and the software is recorded on a recording medium such as a ROM. However, it may be realized by hardware (dedicated circuit).

  In each of the above embodiments, it goes without saying that two or more communication means existing in one apparatus may be physically realized by one medium.

  The database construction device in each of the above embodiments may be, for example, a stand-alone device or a server device in a server / client system. In the latter case, for example, a database may be constructed as an ASP (Application Service Provider).

  In each of the above embodiments, each process or each function may be realized by centralized processing by a single device or a single system, or distributed by a plurality of devices or a plurality of systems. It may be realized by being processed.

  In each of the above embodiments, each component may be configured by dedicated hardware, or a component that can be realized by software may be realized by executing a program. For example, each component can be realized by a program execution unit such as a CPU reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory.

  Moreover, the software which implement | achieves the database construction apparatus in each said embodiment is the following programs. That is, this program has a first table having one or more attribute value sets having m attribute values (m ≧ 3) having different attributes and n items having different attributes (n ≦ m−1). ) One or more for obtaining two or more second tables having one or more attribute value sets having attribute values and one or more attribute value sets having the n attribute values from one or more Web pages An acquisition unit that acquires one or more attribute value sets having the n attribute values from one or more Web pages according to the rules, and appends them to the second table; As a combining unit that combines two or more attribute value sets of the two or more second tables, generates one or more attribute value sets having the m attribute values, and appends to the first table It is a program to make it function.

  Note that the program does not include at least processing that is performed only by hardware.

  The program may be executed by being downloaded from a server or the like, or a program recorded on a predetermined recording medium (for example, an optical disk such as a CD-ROM, a magnetic disk, a semiconductor memory, or the like) is read out. May be executed. Further, this program may be used as a program constituting a program product.

  Moreover, the computer which performs the said program may be single, and plural may be sufficient as it. That is, centralized processing may be performed, or distributed processing may be performed.

  FIG. 18 is a schematic diagram of the computer system 9 that executes the above-described program to realize the database construction apparatus and the like according to the above-described embodiment. The above-described embodiments can be realized by computer hardware and a computer program executed thereon.

  In FIG. 18, a computer system 1100 includes a computer 1101 including a CD-ROM (Compact Disk Read Only Memory) drive 1105, an FD (Flexible Disk) drive 1106, a keyboard 1102, a mouse 1103, and a monitor 1104.

  FIG. 19 is a block diagram of the computer system 1100. In FIG. 19, in addition to the CD-ROM drive 1105 and the FD drive 1106, a computer 1101 is connected to an MPU (Micro Processing Unit) 1111, a ROM 1112 for storing a program such as a bootup program, and the MPU 1111. A RAM (Random Access Memory) 1113 that temporarily stores program instructions and a temporary storage space, a hard disk 1114 that stores application programs, system programs, and data, and an MPU 1111 and a ROM 1112 are mutually connected. And a bus 1115. The computer 1101 may include a network card (not shown) that provides connection to the LAN.

  A program for causing the computer system 1100 to execute the functions of the present invention according to the above-described embodiment is stored in the CD-ROM 1121 or the FD 1122, inserted into the CD-ROM drive 1105 or the FD drive 1106, and transferred to the hard disk 1114. May be. Instead, the program may be transmitted to the computer 1101 via a network (not shown) and stored in the hard disk 1114. The program is loaded into the RAM 1113 when executed. Note that the program may be loaded directly from the CD-ROM 1121, the FD 1122, or the network.

  The program does not necessarily include an operating system (OS) or a third-party program that causes the computer 1101 to execute the functions of the present invention according to the above-described embodiment. The program may include only a part of an instruction that calls an appropriate function (module) in a controlled manner and obtains a desired result. How the computer system 1100 operates is well known and will not be described in detail.

  The present invention is not limited to the above-described embodiments, and various modifications are possible, and it goes without saying that these are also included in the scope of the present invention.

  As described above, the database construction device according to the present invention has an effect that a database can be constructed from a Web page, and is useful as a database system or the like.

1, 2, 5 Database construction device 3, 4 Trademark infringement detection device 101 Company database 102 Product database 103 Unique name database 104 Company product database 105 Company unique name database 106 Product unique name database 107 Company product unique name database 108 Company rule database 109 Product rule database 110 Unique name rule database 111 Knowledge multiplication rule database 112 Company name acquisition unit 113 Product name acquisition unit 114 Unique name acquisition unit 115 Knowledge acquisition unit 116 Rule learning unit 21 Product similarity group code correspondence database 22 Company similarity group code unique name Database 23 Company similar group code unique name database construction unit 31, 401 Database 32 Reception unit 33 Trademark similarity determination unit 34, 4 2 Product type judgment unit 35 Output unit 51 Table storage unit 52 Rule storage unit 53 Score storage unit 54 Dictionary storage unit 55 Acquisition unit 56 Joining unit 57 Rule acquisition unit 331 Trademark character string acquisition unit 332 Name call acquisition unit 333 Name call display unit 334 Name designation accepting unit 335 Similarity search unit 336 Trademark similarity determination result acquisition unit 4021 Product code management means 4022 Code specific information acquisition means 4023 Product similarity determination means

Claims (13)

A first table having one or more attribute value sets having m attribute values (m ≧ 3) having different attributes, and one or more attributes having n attribute values (n ≦ m−1) having different attributes A table storage for storing two or more second tables having a value set;
A rule storage unit that stores one or more rules for acquiring one or more attribute value sets having the n attribute values from one or more Web pages in association with the second table;
In accordance with the rules, an acquisition unit that acquires one or more attribute value sets having the n attribute values from one or more Web pages and appends to the second table;
Combining two or more attribute value sets included in the two or more second tables, generating one or more attribute value sets having the m attribute values, and adding the combined value to the first table; A database construction device provided. In the rule storage unit,
One or more rules for acquiring one or more attribute value sets having the m attribute values from one or more Web pages are stored in association with the first table,
The acquisition unit
The database construction device according to claim 1, wherein one or more attribute value sets having the m attribute values are acquired from one or more Web pages according to the rule, and added to the first table. In the table storage section,
The second table having a unique identifier indicating whether the attribute is unique is stored;
The coupling portion is
Two or more attributes that are attribute value sets that each of the two or more different second tables have, have the same attribute value of the same attribute, and that the attribute is unique by the unique identifier The database construction device according to claim 1, wherein the value set is combined with the attribute value as a key, one or more attribute value sets having the m attribute values are generated, and added to the first table. The coupling portion is
Each of the three or more different second tables has an attribute value set, the attribute value set having two or more identical attribute values of the same attribute is combined with the attribute value as a key, and the m attributes The database construction device according to claim 1, wherein one or more attribute value sets having values are generated and added to the first table. The web page has a web page identifier for identifying the web page,
The acquisition unit
When acquiring the one or more attribute value sets, a Web page identifier of a Web page is also acquired, and the Web page identifier is associated with the one or more attribute value sets in the first table or the second table. The database construction device according to any one of claims 1 to 4, which is added to a table. The acquisition unit
When acquiring the one or more attribute value sets, the score of the attribute value set is also acquired, and the score is added to the first table or the second table in association with the one or more attribute value sets. And
The coupling portion is
Combining two or more attribute value sets included in the second table with a higher score corresponding to a predetermined condition to generate one or more attribute value sets having the m attribute values, The database construction device according to any one of claims 1 to 5, which is added to the first table. A score storage unit that stores one or more scores associated with the web page identifier;
The acquisition unit
When acquiring the one or more attribute value sets, the Web page identifier of the Web page is also acquired, a score corresponding to the Web page identifier is acquired from the score storage unit, and the one or more attribute value sets are included. The database construction device according to claim 5 or 6, wherein the Web page identifier and the score are added to the first table or the second table in association with each other. A dictionary storage unit that stores one or more dictionaries that are a set of one or more attribute values for each attribute;
The acquisition unit
When acquiring the one or more attribute value sets, the attribute value that the attribute value set has and the attribute value of the attribute that is the same attribute as the attribute value, and the degree of coincidence with the attribute value that the dictionary has The database construction apparatus according to claim 6, wherein the database construction apparatus calculates the score and adds the score to the first table or the second table in association with the one or more attribute value sets. A character string including one or more attribute value sets stored in the table storage unit is acquired from the one or more Web pages, and a character string other than the attribute value set among the character strings is used as the rule. The database construction device according to claim 1, further comprising a rule acquisition unit that acquires and accumulates in the rule storage unit. The rule acquisition unit
The database according to claim 9, wherein the acquired one or more rules are stored in the rule storage unit in association with the first table or the second table having the attribute value set used when the rule is acquired. Construction device. The coupling portion is
Two or more attribute value sets of two or more second tables stored in the table storage unit, and two or more of the attribute value sets acquired by the acquisition unit or the attribute value sets generated by the attribute value generation unit 11. The database construction device according to claim 1, wherein the attribute value sets are combined to generate one or more attribute value sets having the m attribute values, and are added to the first table. On the recording medium,
A first table having one or more attribute value sets having m attribute values (m ≧ 3) having different attributes, and one or more attributes having n attribute values (n ≦ m−1) having different attributes Two or more second tables having value sets;
One or more rules for acquiring one or more attribute value sets having the n attribute values from one or more Web pages are stored,
A database construction method performed using an acquisition unit and a combination unit,
The acquisition unit, according to the rule, acquires one or more attribute value sets having the n attribute values from one or more Web pages, and adds to the second table;
The combining unit combines two or more attribute value sets of the two or more second tables, generates one or more attribute value sets having the m attribute values, and adds to the first table A database construction method comprising a combining step. On the recording medium,
A first table having one or more attribute value sets having m attribute values (m ≧ 3) having different attributes, and one or more attributes having n attribute values (n ≦ m−1) having different attributes Two or more second tables having value sets;
One or more rules for acquiring one or more attribute value sets having the n attribute values from one or more Web pages are stored,
Computer
In accordance with the rules, an acquisition unit that acquires one or more attribute value sets having the n attribute values from one or more Web pages and appends to the second table;
Functions as a combining unit that combines two or more attribute value sets of the two or more second tables, generates one or more attribute value sets having the m attribute values, and adds the attribute value sets to the first table Program to let you.

Images