JP5491422B2 - Information search device, information search method, and information search program - Google Patents

Description

  The present invention relates to an information search technique for presenting a document that matches an input search term.

  There are cases where a search keyword used for information search has a plurality of notations. At this time, in addition to the search keyword, another notation having the same meaning is added as a search keyword, which is a search technique for reducing so-called search omission (for example, Patent Document 1 and Non-Patent Document 1).

  According to this search technique, it is possible to present a document having contents that match the user's request without including the keyword input by the user as it is. In particular, when a search keyword is an abbreviation, a more appropriate search result can be obtained by adding a search result with a word that the abbreviation means in addition to a search result with the input abbreviation. For example, consider a case where a search is performed using the keyword “BBQ”. “BBQ” is an abbreviation for “barbecue”. At this time, in addition to the search by the keyword “BBQ”, the search result by the keyword “BBQ” is also presented, so that more documents matching the user's request can be presented.

  Further, a range search technique for performing a search by narrowing down documents included in a specified range is known (for example, Non-Patent Documents 2 and 3). The designated search range includes a time range such as date and time, and a geographical range such as a place and a region. For example, when “August-September 2010” is specified in a range search in which a time range is specified, the date and time expression included in the document, the date when the document is collected, etc. are used. Documents included in “September” can be searched. In general, the user performs a search using a keyword in addition to specifying a search range, and the search system presents a document that satisfies both the specified search range and the specified keyword.

JP 2008-204399 A

Akira Terada, Takenobu Tokunaga, Hozumi Tanaka, "Automatic expansion of abbreviations by using context and character information", Information Processing and Management 40, 2004, January 2004, pp.31-45 Hiroyuki Toda, Yoshihito Yasuda, Yumiko Matsuura, Ryoji Kataoka, “Retrieval of Geographic Information Considering Proximity of Geographic Information and Content Information”, 23rd Annual Conference of Japanese Society for Artificial Intelligence, June 2009 Hideki Kawai, Adam Jatowt, Katsumi Tanaka, Kazuo Kunieda, Keigo Yamada, “On-demand search engine for past and future information from ChronoSeeker Web”, WebDB Forum 2009, November 2009

  A plurality of words having different meanings may be represented by the same abbreviation. For example, the word “ES” is an abbreviation of the word “entry sheet” and also an abbreviation of a plurality of words such as “Employee Satisfaction” and “Engineering Service”.

  The main words that abbreviations mean depend on the search range. For example, consider the case where time is the search range. Speaking of “ES” in a certain period indicates “employee satisfaction”, and it is used in the meaning of “entry sheet” in a certain period. When geography is used as a search range, there are cases where main words that abbreviations differ depending on the region.

  However, in the conventional method, when extracting a word represented by an abbreviation, all document sets are used. As a result, a word other than the meaning intended by the search user is added as a search keyword, resulting in a problem that the accuracy of the search is lowered.

  The present invention has been made in view of the above circumstances, and when an abbreviation is provided as a keyword, an appropriate abbreviation candidate in a designated search range is acquired as an additional keyword for the search, The object is to improve the accuracy of information retrieval based on abbreviations.

  Therefore, the present invention, when performing an information search based on the abbreviation, by using the corresponding word extracted based on the degree of importance from the corresponding word group corresponding to the abbreviation and the specified search range as an additional keyword, the abbreviation Improve search accuracy based on.

As an aspect of the information search apparatus of the present invention, there is an information search apparatus for searching for a document that matches an abbreviation-containing query, from an abbreviation provided as the query and a corresponding word group corresponding to a search range. comprising an additional keyword acquisition means for a corresponding word obtained based on the key of each corresponding word and additional keywords of the query, the keyword, and range search unit for performing information retrieval based on the retrieval range and the additional keyword, the additional keyword acquisition means, if a single search range is provided to calculate a specified range in the primary index showing the main degree of the search range for each respective corresponding word group corresponding to the keyword and the search range , if the major degree to get a higher correspondence word than a threshold value as said additional keywords, a plurality of search ranges are provided, the keyword And the corresponding word in the specified range set consisting of the plurality of search ranges obtained by multiplying the main degrees in the specified ranges indicating the main degrees in the respective search ranges of the corresponding words corresponding to the plurality of search ranges. And a corresponding word having a higher degree of importance than a threshold is acquired as the additional keyword .

An aspect of the information search method of the present invention is an information search method for searching for a document that matches a query including an abbreviation, wherein the additional keyword acquisition means corresponds to a keyword including the abbreviation provided as the query and a search range. A corresponding word acquired from the corresponding word group based on the importance of each corresponding word as an additional keyword of the query, and a step in which range search means performs an information search based on the keyword, the search range, and the additional keyword; have, if single search range is provided, in the step of acquiring the additional keywords, specify showing the principal degree of the search range for each respective corresponding word group corresponding to the keyword and the search range calculating a primary index in the range, the major degree to get a higher correspondence word than a threshold value as said additional keywords, a plurality of search range In the step of acquiring the additional keyword, the degree of importance in the designated range indicating the degree of importance in each search range of each corresponding word in the corresponding word group corresponding to the plurality of search ranges is multiplied in the step of acquiring the additional keyword. And a corresponding word whose degree of importance is higher than a threshold is acquired as the additional keyword .

As described above, when a single search range is provided, the keyword and the specified range in the primary index showing the main degree of the search range for each respective corresponding word group corresponding to the search range is calculated, this Corresponding words whose importance is higher than a threshold are acquired as the additional keywords . In this way, since the corresponding words are extracted based on the threshold value of the majority in the designated range, information search based on the keywords can be performed with a desired search accuracy.

When a plurality of search ranges are provided, which multiplied by the specified range in the primary index showing the main level in each search range of each respective corresponding word group corresponding to the keyword and the plurality of search ranges wherein the specified range set comprising a plurality of search ranges is a major degree in the specified range that indicates the main level of the corresponding word, high correspondence word than the major degree threshold is obtained as the additional keywords. Thereby, even if a plurality of search ranges are pointed out, information search based on the keyword can be performed with desired search accuracy.

  Furthermore, as another aspect of the information search apparatus and information search method of the present invention, in acquiring the additional keyword, a keyword consisting of an abbreviation, a range defining disclosure of the abbreviation, a corresponding word corresponding to the abbreviation, It is preferable to extract a corresponding word group corresponding to the provided keyword and one or a plurality of search ranges from a keyword database in which the degree of corresponding word is stored in advance. According to this aspect, it is possible to efficiently acquire the abbreviations provided and corresponding word groups corresponding to one or a plurality of search ranges.

  Note that the present invention can also be implemented as an information search program that causes a computer to function as each means constituting the information search apparatus.

  According to the above invention, when an abbreviation as a keyword is used for information search, an appropriate abbreviation candidate in the search range specified together with the abbreviation is used as an additional keyword for the search, so information based on the abbreviation Search accuracy is improved.

1 is a block configuration diagram of an information search apparatus according to an embodiment of the invention. The flowchart explaining the process sequence of the information search device which concerns on embodiment of invention. The flowchart explaining the acquisition procedure of the additional keyword by an additional keyword acquisition part. An example of an additional keyword database having a temporal range column. An example of an additional keyword database having a column of geographic ranges.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to the following embodiments.

[Overview]
The information search apparatus 1 according to the present embodiment shown in FIG. 1 is configured such that when a keyword consisting of an abbreviation is used for information search, the correspondence in the search range is determined from the corresponding word group corresponding to the search range specified together with the abbreviation. Corresponding words are extracted based on word importance. The extracted corresponding word is used as an additional keyword for information retrieval.

  In the present embodiment, the information retrieval apparatus 1 includes an additional keyword database 3 (hereinafter referred to as “keywords” including abbreviations, a range defining disclosure of the abbreviations, a corresponding word corresponding to the abbreviation, and a degree of importance of the corresponding word). , Additional keyword DB3). In the information search apparatus 1, a corresponding word group corresponding to the keyword and search range provided from the query input unit 2 is extracted from the additional keyword DB 3, and designation indicating the degree of importance in the search range for each corresponding word of the corresponding word group Calculate the importance in the range. Then, the corresponding word whose degree of importance is higher than the threshold is used as an additional keyword for information search by the range search unit 6. As described above, when an abbreviation is used as a search keyword, corresponding words represented by the abbreviation are narrowed down by a designated search range, and search accuracy is improved.

[Configuration of Information Retrieval Device 1]
As shown in FIG. 1, the information search apparatus 1 includes a query input unit 2, an additional keyword DB 3, an additional keyword acquisition unit 4, a document database 5 (hereinafter, document DB 5), a range search unit 6, and a search result output unit 7. Is provided.

  The functional unit and DB of the information search apparatus 1 are realized by hardware resources of a server (computer). That is, the information search apparatus 1 includes hardware resources related to a computer such as a CPU, a storage device (for example, a hard disk device), and an I / O device (for example, a communication device). The functional unit and DB are implemented by these hardware resources cooperating with software resources (OS, applications, etc.).

  The query input unit 2 provides the query received from the user via the network to the additional keyword acquisition unit 4. The query is an information request provided to the information search apparatus 1 and includes a keyword consisting of an abbreviation and a search range. Examples of the search range include time, geography, language, domain, field, and the like. The query input unit 2 is provided on a user interface exemplified by a Web interface.

  As illustrated in FIGS. 4 and 5, the additional keyword DB 3 includes a keyword indicating the abbreviation w, a range t that defines the disclosure of the abbreviation w, a corresponding word e corresponding to the abbreviation w, and a degree of importance of the corresponding word e. This is a database that stores c in advance. The degree of importance c is a score indicating the likelihood of the corresponding word e with respect to the abbreviation w, and the higher the value, the more suitable the corresponding word. The degree of importance c is obtained, for example, by the well-known “Score” calculation method disclosed in Non-Patent Document 1.

  The additional keyword DB illustrated in FIG. 4 is an example of a database when the search range is a time range. The additional keyword DB illustrated in FIG. 5 is an example of a database when the search range is a geographic range. The corresponding word e corresponding to the abbreviation w in the range t is extracted by using a document set that matches the range t. Acquisition of a document set that conforms to the range t can be performed by the same method as the method for specifying a document that conforms to the designated search range by the range search unit 6 described later.

  The additional keyword acquisition unit 4 extracts a corresponding word group corresponding to the keyword and the search range specified by the query input unit 2 from the additional keyword DB 3, and for each corresponding word, a specified range indicating the degree of importance of the corresponding word in the search range Calculate medium importance. Then, a corresponding word having a higher degree of importance than a threshold is acquired as an additional keyword.

  Further, when a plurality of search ranges are designated by the query input unit 2, the additional keyword acquisition unit 4 searches the keyword provided from the query input unit 2 and a corresponding word group corresponding to the plurality of search ranges from the additional keyword DB 3. Pull out. Next, for each corresponding word in the corresponding word group, the degree of importance in the designated range indicating the degree of importance of the corresponding word in each search range is calculated. Next, the product of the degrees of importance in the designated ranges in the respective search ranges is calculated as the degree of importance in the designated ranges indicating the degrees of importance of the corresponding words in the designated range set consisting of the plurality of search ranges. Then, a corresponding word having a higher degree of importance than a threshold is acquired as an additional keyword.

  The specific calculation of the majority in the designated range of the corresponding word by the additional keyword acquisition unit 4 when “single search range is designated” and “when multiple search ranges are designated” will be described respectively.

(Calculation of the importance in the specified range when a single search range is specified)
The search range specified by the user may be different from the unit of the range in the database. For example, in the database example of FIG. 4, the time unit is set to one year, but the user designates from July 2008 to December 2009. In order to cope with such a case, the degree of importance in the designated range, which is the degree of importance of the corresponding word in the search range designated by the user, is calculated.

  The degree of significance sc (w, s, e) of the corresponding word e in the specified range (search range) s for the abbreviation w can be calculated by the following equation (1).

  Here, t represents a range, and T represents a set of ranges t including the specified range s. If the specified range s is from July 2008 to December 2009, T is 2008 and 2009. λ represents a ratio in which the specified range s is included in the range t. Considering the previous example, if t is 2008, the specified time includes the half year from July to December, so the ratio λ is 0.5. On the other hand, if the range t is 2009, the specified range is included from January to December, so the ratio λ is 1. c (w, t, e) represents the importance of the corresponding word e in the range t with respect to the keyword w. Then, a corresponding word e having a degree of importance sc in the designated range that is higher than a preset threshold value is determined as an additional keyword.

  The threshold is arbitrarily set. That is, the threshold is set low when it is desired to add more keywords and present more search results, and the threshold is set high when it is desired to increase the accuracy by reducing unrelated search results.

(Calculation of the importance in the specified range when multiple search ranges are specified)
When a plurality of search ranges are specified as search conditions, the priority of the corresponding words in the specified range set of the plurality of search ranges is multiplied by multiplying the degrees of importance in the specified ranges calculated in each search range. Is calculated.

That is, when n search ranges are specified, each specified range (search range) is represented by S _i (i = 1, 2, 3,..., N), and a set of the specified ranges S _i is S. . The degree of significance sc (w, S, e) in the designated range of the corresponding word e in the designated range set S for the abbreviation w can be calculated by the following equation (2). Then, the corresponding word e in which the value of the degree of importance sc in the specified range is higher than a preset threshold value is determined as an additional keyword. The threshold value is also arbitrarily set according to the above-mentioned purpose.

(Additional keyword acquisition procedure by additional keyword acquisition unit 4)
The additional keyword acquisition procedures S41 to S47 by the additional keyword acquisition unit 4 will be described with reference to FIG. Here, an additional keyword acquisition procedure in a case where the specified range s is specified as the search range will be described.

  (S41) A keyword (abbreviation) and a specified range s are input from the query input unit 2.

  (S42) A line having a range t including the keyword and the specified range s is searched in the additional keyword DB3.

  (S43) If the row exists, the process proceeds to S44. If the row does not exist, the process ends.

  (S44) The degree of importance sc in the designated range of the corresponding word e is calculated from the line acquired in S42 by the calculation according to the equation (1).

  (S45) If the calculated degree of importance sc in the specified range is greater than or equal to the threshold, the process proceeds to S46. If the degree of importance sc in the specified range is not greater than or equal to the threshold value, the process proceeds to S47.

  (S46) The corresponding word / phrase e having a degree of importance sc in the specified range that is equal to or greater than a threshold is acquired as an additional keyword.

  (S47) The importance sc in the designated range is calculated in S44 for all the corresponding words e in the row having the range t including the keyword and the designated range s in the additional keyword DB3.

  The additional keyword acquired by the above procedure is provided to the range search unit 6.

  The document DB 5 is a database that stores previously collected documents. For example, it is a database constructed by recording a document published on the Web by a known information collecting means such as a crawler. If the content of the document is an information source that is sequentially updated, such as a website on the Internet, when the document is updated, it is considered that a new document has been created and is collected by the information collecting means. Stored in the document DB. It should be noted that existing Web page search engines may be used in place of the document DB 5.

  The range search unit 6 searches the document DB 5 for documents that match the input keyword, documents that match the input search range, and documents that match the additional keyword obtained by the additional keyword acquisition unit 4. A well-known search technique is applied to search for the document.

  The search result output unit 7 outputs a document obtained by the search by the range search unit 6 as a search result. Similar to the query input unit 2, the search result output unit 7 is provided on the user interface.

[Description of Search Processing Procedure by Information Search Device 1]
The procedure of document search by the information search apparatus 1 will be specifically described with reference to FIG.

  S1: The query input unit 2 provides the keyword (abbreviation) and search range received as a query from the user to the additional keyword acquisition unit 4.

  The search range includes, for example, time, geography, language, domain, and field. When the search range is time, “July 2008-December 2009” is exemplified as the search range. If the search range is geography, the search range is “Northeast coordinates latitude 35 degrees 00 minutes 00 seconds, longitude 139 degrees 00 minutes 00 seconds to southwest coordinates latitude 34 degrees 50 minutes 00 seconds, longitude 138 degrees 50 minutes 00 seconds. Is exemplified. When the search range is a language, a search range query such as “Japanese” may be used. In addition, domains, fields, etc. are exemplified as the search range.

  S2: The additional keyword acquisition unit 4 determines whether the search range provided from the query input unit 2 is included in the additional keyword DB3.

  In S2, specifically, the above-described steps S41 to S43 described with reference to FIG. 3 are executed. For example, it is determined whether or not the search range is included in the additional keyword DB 3 shown in FIGS. When the search range is included in the additional keyword DB 3, the process proceeds to S3. On the other hand, when the search range is not included in the additional keyword DB 3, the process proceeds to S7.

  S3: The additional keyword acquisition unit 4 extracts a corresponding word group in a range corresponding to the search range provided from the query input unit 2 from the additional keyword DB3.

  S4: The additional keyword acquisition unit 4 calculates the majority in the designated range of each corresponding word of the corresponding word group acquired in S3, and acquires the corresponding word having the calculated majority higher than the threshold as an additional keyword.

  In S4, specifically, the above-described steps S44 to S47 described with reference to FIG. 3 are executed. When the search range provided from the query input unit 2 is single, the degree of importance in the designated range of the corresponding word is calculated by the calculation based on the formula (1). On the other hand, when there are a plurality of search ranges provided from the query input unit 2, the degree of importance in the designated range of the corresponding word is calculated by the calculation based on the formula (2). Then, a corresponding word / phrase whose degree of importance in the designated range is equal to or greater than a threshold is acquired as an additional keyword.

  S5: The additional keyword acquisition unit 4 supplies the range search unit 6 with a query obtained by adding the additional keyword acquired in S4 to the keyword and search range provided in S1.

  S6: The range search unit 6 searches the document DB 5 for a document that matches the keyword input in S1, a document that matches the search range input in S1, and a document that matches the additional keyword obtained in S5.

  S7: The additional keyword acquisition unit 4 acquires the keyword by using the search range input in S1 for a search engine that executes a range search according to the prior art exemplified in Non-Patent Documents 2 and 3, and uses this as an additional keyword. To do. And the query which added the additional keyword acquired by this step S7 with respect to the keyword and search range which were provided by S1 is provided to the range search part 6. FIG.

  S8: The range search unit 6 searches the document DB 5 for a document that matches the keyword input in S1, a document that matches the search range input in S1, and a document that matches the additional keyword obtained in S7.

  S9: The search result output unit 7 outputs the document acquired by the search in S6 or S8 as a search result.

[Effect of this embodiment]
As described above, according to the information search apparatus 1 of the present embodiment, when an abbreviation and a search range are used as search queries, search omissions are reduced and search accuracy is improved. In addition, since the corresponding words represented by the abbreviation are narrowed down by the range specified when the abbreviation is specified as the search query, the abbreviation and the document unrelated to the range are excluded.

  In particular, since the additional keyword DB 3 is provided in the present embodiment, it is possible to efficiently acquire the corresponding word group corresponding to the abbreviation and the single or plural search ranges.

  In addition, for each corresponding word in the corresponding word group corresponding to the abbreviation and the search range, a degree of importance in the designated range indicating the degree of importance in the search range is calculated, and a corresponding word having a degree of importance higher than a threshold is added to the search query. It is acquired as an additional keyword. As described above, the corresponding words are extracted based on the threshold value of the degree of importance in the designated range, so that information search based on the abbreviations can be performed with desired search accuracy.

  Further, when a plurality of search ranges are designated, the abbreviations and the majorities in the designated ranges indicating the majorities in the respective search ranges of the corresponding words of the corresponding word group corresponding to the plurality of search ranges are multiplied. The degree of importance in the designated range indicating the degree of importance of the corresponding word in the designated range set consisting of the plurality of search ranges. Then, a corresponding word having a degree of importance higher than a threshold value is acquired as an additional keyword to the search query. Therefore, even when a plurality of search ranges are pointed out, information search based on the abbreviation can be performed with desired search accuracy.

  Further, in the present embodiment, the document DB 5 storing information collected in advance is provided, so that a document that matches the abbreviation, the search range, and the additional keyword can be quickly acquired.

  The present invention is not limited to the above-described embodiment, and various modifications and applications can be made within the scope of the claims.

[Aspect as Program of the Present Invention]
The present invention can implement the present invention by configuring a part or all of the functions of each functional unit of the information search apparatus 1 of the above embodiment by a computer program, and executing the program by the computer. In addition, a computer-readable recording medium such as an FD (Floppy (registered trademark) Disk), an MO (Magneto-Optical disk), or a ROM (Read Only Memory) can be read by the computer. , Memory card, CD (Compact Disk) -ROM, DVD (Digital Versatile Disk) -ROM, CD-R, CD-RW, HDD, SSD, removable disk, etc. Is possible. It is also possible to provide the above program through a network such as the Internet or electronic mail.

1 ... information retrieval device 3 ... additional keyword DB (keyword database)
4. Additional keyword acquisition unit (additional keyword acquisition means)
6. Range search unit (range search means)

Claims (5)

An information retrieval device for retrieving documents that match a query including an abbreviation,
An additional keyword acquisition means that uses a keyword consisting of abbreviations provided as the query and a corresponding word acquired from a corresponding word group corresponding to a search range based on the importance of each corresponding word as an additional keyword of the query;
Range search means for performing information search based on the keyword, the search range, and the additional keyword;
With
The additional keyword acquisition means includes
If a single search range is provided to calculate a specified range in the primary index showing the main degree of the search range for each respective corresponding word group corresponding to the keyword and the search range, the major degree threshold Higher corresponding word as the additional keyword,
In the case where a plurality of search ranges are provided, the keywords and the corresponding word groups corresponding to the plurality of search ranges are multiplied by the plurality of degrees in the designated range indicating the degrees of importance in each search range of the corresponding words. A degree of importance in the designated range indicating the degree of importance of the corresponding word in the designated range set consisting of the search ranges, and a corresponding word having a degree of importance higher than a threshold value is acquired as the additional keyword. Information retrieval device. A keyword database that pre-stores keywords indicated by abbreviations, ranges defining disclosure of the abbreviations, corresponding words corresponding to the abbreviations, and the majority of the corresponding words;
2. The information search according to claim 1 , wherein the additional keyword acquisition unit extracts a corresponding word group corresponding to the keyword and the search range from the keyword database based on the provided keyword and one or a plurality of search ranges. apparatus. An information retrieval method for retrieving documents that match a query including an abbreviation,
The additional keyword acquisition means sets the corresponding word acquired based on the majority of each corresponding word from the corresponding word group corresponding to the keyword and the search range consisting of the abbreviation provided as the query as an additional keyword of the query;
A range search means for performing an information search based on the keyword, the search range and the additional keyword;
Have
If a single search range is provided, wherein in the step of acquiring additional keywords, the keyword and the specified range in the primary index showing the main degree of the search range for each respective corresponding word group corresponding to the search range To obtain a corresponding word having a degree of importance higher than a threshold as the additional keyword,
When a plurality of search ranges are provided, in the step of obtaining the additional keyword, the key in the specified range indicating the degree of importance in each search range of each corresponding word of the corresponding word group corresponding to the keyword and the plurality of search ranges The degree multiplied by the degree is used as the degree of importance in the designated range indicating the degree of importance of the corresponding word in the designated range set composed of the plurality of search ranges, and the corresponding word having the degree of importance higher than the threshold is acquired as the additional keyword. information retrieval wherein the <br/> that. In the step of acquiring the additional keyword, the keyword is provided from a keyword database that stores in advance a keyword indicated by an abbreviation, a range defining disclosure of the abbreviation, a corresponding word corresponding to the abbreviation, and a degree of importance of the corresponding word. 4. The information search method according to claim 3 , wherein a corresponding word group corresponding to the keyword and one or a plurality of search ranges is extracted. An information search program for causing a computer to function as each means constituting the information search apparatus according to claim 1 .

Images