JP2005275714A - Information retrieval apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information retrieval apparatus by which a retrieval result is obtained corresponding to a change in the meaning of a word with the lapse of time. <P>SOLUTION: The control section 101 of the information retrieval apparatus 100 makes a keyword extract section 103 extract a keyword regarding an occupation in a sentence collected from a site on the specified Internet by a Web information collection section 102 and to calculate a weight being the degree of association of the keyword with the occupation to write the weight into a keyword history DB (Database) 104. The control section 101 makes an index generation section 106 totalize the weight for every name of the occupation based on the keyword history DB 104 and to store the weight on an Internet index DB 107. The control section 101 makes a retrieval processing section 108 read the names of the occupations from an occupation DB 110 upon receiving the keyword specified by a user, and to return the list of the names of the occupations taking the totalized weight in the Internet index DB 107 into account. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

   The present invention relates to an information retrieval apparatus for keyword retrieval via the Internet.

  There are search sites on the Internet where general users can specify keywords and obtain words related to specific fields such as computer terms and economic terms as search results. The administrator of such a search site tunes the search site system in order to improve the search accuracy. However, in the conventional technology, this tuning is performed manually based on the frequency of use of keywords on the search site, indications from users, or advice from experts who are familiar with the specific field to be searched. It was. For this reason, in the field where the meaning of the word changes with the passage of time, there is a problem that the word cannot appear in the search result unless the user uses an old keyword because the tuning cannot catch up.

In addition, what is described in patent document 1 is known as a conventional search device. The technique described in Patent Document 1 adds a similar keyword to a search target word in a search device, and the search site cannot cope with a change in the meaning of the word itself with the passage of time to be solved by the present invention. It does not solve the problem.
Japanese Patent Laid-Open No. 11-272706

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information search apparatus capable of obtaining a search result corresponding to a change in the meaning of a word itself with the passage of time.

  The present invention has been made to solve the above-mentioned problems. The invention of claim 1 is a word database in a storage area for storing a word and a keyword related to the word when a keyword is received from a client computer. An information search device that reads out a word corresponding to the received keyword from the client and transmits the word to the client computer, the weight representing a word, a keyword related to the word, and a high degree of association between the word and the keyword A keyword history database that stores information, an information collection unit that accesses a plurality of sites on the Internet designated in advance to obtain information on sentences posted on the site, and information on sentences obtained by the information collection unit Extracting a keyword related to a word in the keyword history database; A keyword extraction unit that calculates a weight, writes the word, the keyword, and the weight into the keyword history database, reads data from the keyword history database, and combines the word and the keyword as a key When the weight is obtained, the weight is obtained, and when the keyword is received from the client computer, the word generator, the keyword, the index generation unit for writing the weighted weight to the Internet index database in the storage area, and the word Read a word corresponding to the received keyword from the database, determine a relevance between the received keyword and the word corresponding to the received keyword based on the aggregated weight in the Internet index database, A word corresponding to the received keyword, a retrieval processing unit for transmitting said relevance to the client computer is an information retrieval apparatus comprising: a.

  The invention according to claim 2 is the information search device according to claim 1, wherein the keyword extraction unit and the index generation unit perform processing every predetermined period. .

  The invention according to claim 3 is the information search apparatus according to claim 1 or 2, further comprising a keyword history database maintenance unit that deletes data that has passed a predetermined time from the keyword history database. Furthermore, it comprises.

  A fourth aspect of the present invention is the information search device according to any one of the first to third aspects, wherein the keyword history database maintenance unit performs processing every predetermined period. Is.

  According to a fifth aspect of the present invention, when a keyword is received from a client computer, a word corresponding to the received keyword is read from a word database in a storage area for storing a word and a keyword related to the word, and the client The computer of the information retrieval device to be transmitted to the computer, a process of accessing a site on the Internet designated in advance to obtain information on the text posted on the site, a word, a keyword related to the word, and the word Stored in a keyword history database in a storage area for storing a weight representing a high degree of relevance with the keyword, an address of a site where the keyword was posted, and an extraction date and time of the keyword from the site. Extract keywords from the sentence information A weight that is a degree of association between the extracted keyword and the word, and the word, the extracted keyword, the weight, the site address, and the date and time when the keyword was extracted. A process of writing to the keyword history database, reading data from the keyword history database, obtaining a weight obtained by summing the weights using the word and the keyword together as a key, and obtaining the word, the keyword, and the sum A process of writing the weights to the Internet index database in the storage area, and when a keyword is received from the client computer, the word corresponding to the received keyword is read from the word database, the received keyword, and the reception Corresponding to the selected keyword For executing a process of transmitting a word corresponding to the received keyword and the degree of association to the client computer. It is a program.

  According to the first or fifth aspect of the present invention, the information search device rewrites the Internet index database with the latest contents based on the Internet site. There is an effect that a search result corresponding to the change can be obtained.

  According to the invention of claim 2, since the Internet index database is not updated for a long time, it is possible to prevent the search result of the information search device from reflecting the appearance of a new word or a change in the meaning of the word.

  According to the invention of claim 3, since old data is deleted from the keyword history database, old data is also deleted from the Internet index database, and the meaning of the old word remains in the search result of the information search device. There is an effect that can be suppressed.

  According to the invention of claim 4, since old data is periodically deleted from the keyword history database, there is an effect that the meaning of the old word remains in the search result of the information search device can be more reliably suppressed.

  First, the basic concept in the present embodiment will be described with reference to FIG. FIG. 2 shows an outline of an information search apparatus regarding occupation information of a site that introduces occupations on the Internet. In this figure, the information retrieval apparatus periodically accesses a plurality of Web sites that provide information on occupations via a network, and collects text data including information on occupations. Next, the information retrieval apparatus analyzes the collected sentence data and extracts a job name included in the sentence data and a keyword when searching for the occupation. Then, the information search apparatus calculates the degree of association with the occupation name based on the appearance frequency on each Web site for the extracted keyword, and registers it in the Internet index DB (Data Base). The user of the Internet uses the information search apparatus of FIG. 2 to search for a job type that the user is interested in based on a keyword such as “care support”, and the “care manager” or the like having a high degree of relevance to the keyword. You can know specific job types.

FIG. 1 shows a configuration of an information search apparatus according to this embodiment. An information search apparatus 100 in FIG. 1 is a computer connected to the Internet. In this embodiment, an information search service related to occupation is provided to general Internet users. The control unit 101 is a control function of the information search apparatus 100, and details will be described later. The Web information collection unit 102 accesses another Web site via the Internet, obtains information about the occupation, and outputs the information about the occupation to the keyword extraction unit 103.
The keyword extraction unit 103 inputs information related to occupations from the Web information collection unit 102, extracts occupation names and occupation related keywords, and writes them to the keyword history DB 104.

  The keyword history DB 104 is a database having the format shown in FIG. The keyword history DB 104 is a database that includes occupation names, keywords relating to occupations, URLs (Universal Resource Locators) of websites where the keywords are registered, keyword registration dates and times, and weights that indicate the degree of association of the keywords with respect to occupations. .

  The website information storage unit 109 stores information related to a site on the Internet, which is an information source related to the occupation of the information search apparatus 100. The Web site information storage unit 109 stores data in the format shown in FIG. 4, where the URL is the URL of the site on which information related to occupation is posted, and the standard / non-standard type indicates whether the site is described in a standard format. The format information is information relating to the format when the site is written in a standard format. The keyword history DB maintenance unit 105 receives an instruction from the control unit 101 and periodically deletes old data in the keyword history DB 104. The index generation unit 106 reads the keyword history DB 104 in response to an instruction from the control unit 101, totals the read data, and writes it in the Internet index DB 107.

The Internet index DB 107 is a database having the format shown in FIG. The Internet index DB 107 is a database that includes occupation names, keywords related to occupations, and aggregated weights that are the aggregated results of the weights of the keyword history DB 104.
The search processing unit 108 inputs a keyword designated by the user from the control unit 101, accesses the occupation DB 110 and the Internet index DB 107, and receives information about the occupation related to the keyword designated by the user via the control unit 101. To provide.
The occupation DB 110 is a database represented in FIG. The occupation DB 110 is a database provided in the information search apparatus 100 in advance for providing occupation information to general Internet users, and includes occupation names, keywords relating to occupations, and explanations of occupations.

Next, the flow of processing in the present embodiment will be described with reference to the drawings.
In the present embodiment, as a preparation, the administrator of the information search apparatus 100 selects a site that provides information related to occupation and designates it as the information search apparatus. Next, the information search device 100 is a process of periodically accessing a site designated by the administrator of the information search device, obtaining information on occupations, extracting keywords from this information, and storing the keywords in the keyword history DB 104. A keyword extraction process is performed. Further, the information search apparatus 100 performs maintenance processing of the keyword history DB 104 every predetermined period. Next, the information search apparatus 100 calculates the use frequency of the keyword used for each occupation from the keyword obtained by the keyword extraction process, performs weighting, and registers the keyword and its weight in the Internet index DB 107. A certain index generation process is performed. Next, when the information search apparatus 100 receives a search request related to occupation from the user via the Internet, the information search apparatus 100 performs a search process for providing the search result to the user. Hereinafter, these processes will be described.

<Website selection>
First, the administrator of the information search apparatus 100 selects other sites that provide information on occupations. Based on information obtained from this site, the information search apparatus 100 stores keywords related to occupation in the keyword history DB 104 inside the information search apparatus. FIG. 3 shows the site selection procedure. The administrator of the information search apparatus 100 uses the administrator's personal computer to search for other sites that provide information about the occupation, and the occupation name is posted using a service such as a search engine on the Internet. When the site is checked and a site providing information on occupation is found, the URL of the site is input to the information search apparatus 100.

The control unit 101 of the information search apparatus 100 receives data input from the administrator, and writes the input URL in the Web site information storage unit 109 as shown in FIG.
In addition, the administrator of the information retrieval apparatus 100 checks whether the site displays information in a fixed format, or does not have a specific format. Format information including information on the display format of the site is created, and this format information is input to the information search apparatus 100. The control unit 101 of the information search apparatus 100 receives data input from the administrator, and stores and writes the input format information and “standard” in the website information storage unit 109.

Further, when the site does not have a fixed format, the administrator of the information search device inputs to the information search device 100 that the site does not have a fixed format. The control unit 101 of the information search apparatus 100 accepts data input from the administrator, and stores and writes “non-standard” in the Web site information storage unit 109.
Thereafter, the information search apparatus 100 accesses the Web site indicated by the URL as shown in FIG. 3 and obtains data including information on occupations as data in HTML (Hyper Text Markup Language) format or XML (extensible Markup Language) format. This process will be described in the keyword extraction process.

<Keyword extraction process>
The information search apparatus 100 performs keyword extraction processing according to the processing flow shown in FIGS. First, the control unit 101 of the information retrieval apparatus 100 periodically accesses a website designated by the administrator of the information retrieval apparatus 100 and collects data in HTML format or XML format including occupation information. A Web information collection request is output to the unit 102. The Web information collection unit 102 receives a request from the control unit 101, reads the URL of the Web site to be accessed from the Web site information storage unit 109, accesses the Web site via the Internet, and includes the HTML format including occupation information Alternatively, page information that is XML format data is received (step S01 in FIG. 5). Next, the Web information collection unit 102 outputs the received page information and the URL read from the Web site information storage unit 109 to the keyword extraction unit 103.

The keyword extraction unit 103 inputs data from the Web information collection unit 102 and refers to the Web site information storage unit 109 to check whether the URL input from the Web information collection unit 102 has a fixed format.
When the Web site from which the page information input from the Web information collecting unit 102 is acquired has a fixed format (S02 in FIG. 5, U01 in FIG. 6, and Step U01 in FIG. 7), where the occupation is located in the page information It is clear beforehand whether it is a keyword for explaining or where a sentence for explaining the occupation exists. For this reason, the keyword extraction unit 103 extracts keywords related to occupations from the page information. Alternatively, for a sentence such as an explanation related to occupation in the page information, the keyword extraction unit 103 performs a morphological analysis process on the item to extract a keyword related to occupation.

  Next, the keyword extraction unit 103 writes the extracted keyword, the URL input from the Web information collection unit 102 and the current date and time in the keyword history DB 104. The keyword extraction unit 103 associates the current date and time with the registration date and time in the keyword history DB 104. Here, even if the same keyword or URL is already registered in the keyword history DB 104, the keyword extracting unit 103 regards the occupation name, the keyword, the URL, and the registration date / time as different data. To the keyword history DB 104.

Next, the keyword extraction unit 103 calculates the weight of the extracted keyword. The keyword extraction unit 103 sets 100 points as the weight of the keyword in the keyword history DB 104 when the extracted keyword is listed as a similar occupation name in the page of the Web site. Further, the keyword extraction unit 103 sets 1 point as the weight of the keyword in the keyword history DB 104 when the extracted keyword appears in the commentary on the page of the Web site.
However, if data that matches all of the occupation name, the keyword, the URL, and the registration date / time has already been registered in the keyword history DB 104, the keyword extraction unit 103 accumulates and adds the weight of the keyword history DB 104. And set.
According to the weighting by the keyword extraction unit 103, in FIG. 7, for example, “care worker”, “home helper”, and “service provider” are each 100 points, and “shift creation” and “user” ”,“ Care manager ”,“ service ”,“ coordination ”,“ visit ”,“ care ”,“ plan ”,“ documents ”,“ maintenance ”,“ helper ”, “Interview”, “recruitment”, “education”, “car”, and “tour” are 5 points each.

  If the Web site from which the page information input from the Web information collection unit 102 is acquired is in an atypical format, the keyword extraction unit 103 generates plain text data by removing HTML tags from the input page information. (T01 in FIG. 6, step T01 in FIG. 8). Next, the keyword extraction unit 103 executes morphological analysis processing on the plaintext data, and divides the plaintext data into words (T02 in FIG. 6 and step T02 in FIG. 8).

  Further, the keyword extraction unit 103 includes symbols such as “.” And “,” in plaintext data, particles such as “ga”, “no”, “wo”, auxiliary verbs such as “da”, “da”, “ Also, connectives such as “,” adverbs such as “almost”, conjuncts such as “so-called” and “some”, or pronouns such as “that” and “this” are removed, leaving only the keywords (T03 in FIG. 6). Step T03 in FIG. 8). Then, the keyword extraction unit 103 writes each keyword in the plain text data in the keyword history DB 104 as in the case where the above-described Web site has a fixed format (T04 in FIG. 6 and step T04 in FIG. 8).

<Keyword history DB maintenance processing>
The information retrieval apparatus 100 periodically repeats the keyword extraction process, so that data related to new occupations is always accumulated in the keyword history DB 104. However, if the information search apparatus 100 only accumulates data, old data remains in the data in the keyword history DB 104. Therefore, the control unit 101 of the information search apparatus 100 outputs a maintenance request to the keyword history DB maintenance unit 105 every predetermined period in order to remove old data from the keyword history DB 104.

  The keyword history DB maintenance unit 105 inputs a maintenance request from the information search apparatus 100, and deletes data in the keyword history DB 104 whose registration date is older than a predetermined time limit (S03 in FIG. 5). By this processing, for example, as shown in FIG. 10, data before 2002/12/31 in the keyword history DB 104 is deleted. Therefore, occupation names and keywords that have become obsolete and no longer used by anyone are deleted from the keyword history DB 104. . Also, in the keyword extraction process, a keyword already stored in the keyword history DB 104 is regarded as different data if the registration date / time is different, and is registered again in the keyword history DB 104. Therefore, it is generally used for a long time. Occupation names and keywords remain in the keyword history DB 104.

<Index generation processing>
The information search apparatus 100 generates a total result of the keyword history DB 104 for each predetermined period and stores the total result in the Internet index DB 107. For this purpose, the control unit 101 outputs an index generation request to the index generation unit 106 every predetermined period.
The index generation unit 106 receives an index generation request from the control unit 101, and once deletes all data in the Internet index DB 107. Then, the index generation unit 106 reads the keyword history DB 104, obtains weights obtained by adding weights using the occupation names in the keyword history DB 104 and the keywords as keys as shown in FIG. And the keyword and the total weight are written (step S04 in FIG. 5).

<Search process>
Now, the information search apparatus 100 receives the keyword “elderly” designated by the user via the Internet, and accepts a search request for an occupation name strongly related to the keyword (step V01 in FIG. 9). In response to this request, the control unit 101 of the information search apparatus 100 outputs the received keyword to the search processing unit 108 and requests a search for an occupation name having a strong relationship with the keyword.

  The search processing unit 108 inputs data from the control unit 101, and generates the occupation list shown in FIG. 13 in the storage area. Then, the search processing unit 108 accesses the occupation DB 110 shown in FIG. 12 to determine whether or not a keyword corresponding to “elderly” appears for each occupation name, and “elderly” appears in the commentary. Check whether or not. The search processing unit 108 gives points to each occupation name while performing this check, and determines that the occupation name having a high point is highly related to the “elderly person” designated by the user. In this processing, for example, “elderly” is not supported as a keyword for “care worker” in the occupation DB 110, but “elderly” is included in the commentary. In this case, the search processing unit 108 gives 50 points to “care worker” and stores “care worker” and 50 points in the occupation list. If points are already given to the “care worker” in the occupation list, the search processing unit 108 adds 50 points and writes them in the occupation list.

  In addition, when “elderly person” corresponds as a keyword to the data of “care worker” in the occupation DB 110, the search processing unit 108 gives 500 points to “care worker”. "Care worker" and 500 points are stored in the occupation list. If points are already given to the “care worker” in the occupation list, the search processing unit 108 adds 500 points to the “care worker” and writes it in the occupation list. The search processing unit 108 performs the same process on other data in the occupation DB 110.

Further, the search processing unit 108 accesses the Internet index DB 107 and checks whether “elderly” is present as a keyword in the Internet index DB 107. When “elderly” is present as a keyword in the Internet index DB 107, the search processing unit 108 reads out the occupation name corresponding to “elderly” and the total weight from the Internet index DB 107 and matches it with the occupation list. Check if the same occupation name is included.
When it is confirmed that the occupation name read from the Internet index DB 107 is not included in the occupation list, the search processing unit 108 regards the total weight as a point and determines the occupation name and the point as the occupation list. Add to
In addition, when the search processing unit 108 detects that the occupation name read from the Internet index DB 107 is included in the occupation list, the search processing unit 108 reads the data point in the corresponding occupation list from the Internet index DB 107. The totaled weights are added and written (step V02 in FIG. 9).
Next, the search processing unit 108 reads out the occupation list from the storage area, sorts the job list in descending order, and outputs it to the control unit 101 (step V03 in FIG. 9). The control unit 101 inputs data from the search processing unit 108 and transmits the occupation list to the request source via the Internet (step V04 in FIG. 9).

  As described above, the embodiments of the present invention have been described in detail with reference to the drawings, but the specific configuration is not limited to these embodiments, and includes design changes and the like within a scope not departing from the gist of the present invention. It is. For example, the information retrieval apparatus in the present embodiment is intended to provide information on occupations, but may be applied to information intended to provide information on computer terms, economic terms, terms related to current affairs information, etc. good.

It is a block diagram showing the structure of the information search device in embodiment of this invention. It is a figure showing the fundamental view of the information search device in embodiment of this invention. It is a figure showing the selection procedure of the Web site which the administrator of the information retrieval apparatus in embodiment of this invention performs. It is a figure showing the content of the Web site information storage part 109 of the information search device in embodiment of this invention. It is a flowchart of the information search device in the embodiment of the present invention. It is a flowchart of the information search device in the embodiment of the present invention. It is a flowchart of the information search device in the embodiment of the present invention. It is a flowchart of the information search device in the embodiment of the present invention. It is a flowchart of the information search device in the embodiment of the present invention. It is a figure showing keyword history DB of the information search device in an embodiment of the invention. It is a figure showing the creation method of internet index DB of the information search device in embodiment of this invention. It is a figure showing occupation DB of the information search device in embodiment of this invention. It is a figure showing the occupation list of the information search device in embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 ... Information retrieval apparatus 101 ... Control part 102 ... Web information collection part 103 ... Keyword extraction part 104 ... Keyword history DB
105 ... Keyword history DB maintenance unit 106 ... Index generation unit 107 ... Internet index DB
108 ... Search processing unit 109 ... Web site information storage unit 110 ... Occupation DB

Claims (5)

An information retrieval device for reading a word corresponding to the received keyword from a word database in a storage area for storing a word and a keyword related to the word when the keyword is received from a client computer and transmitting the word to the client computer. There,
A keyword history database storing a word, a keyword related to the word, and a weight representing a high degree of association between the word and the keyword;
An information collecting unit for accessing a plurality of sites on the Internet designated in advance and obtaining information on sentences posted on the sites;
A keyword related to a word in the keyword history database is extracted from sentence information obtained by the information collecting unit, a weight of the keyword is calculated, and the word, the keyword, and the weight are written in the keyword history database. A keyword extractor;
Data is read from the keyword history database, the weight obtained by adding the word and the keyword together as a key to obtain the weight is obtained, and the word, the keyword, and the collected weight are stored in the Internet in the storage area. An index generator for writing to the index database;
When a keyword is received from the client computer, a word corresponding to the received keyword is read from the word database, and a degree of association between the received keyword and the word corresponding to the received keyword is stored in the Internet index database. A search processing unit for determining a word corresponding to the received keyword and transmitting the degree of association to the client computer;
An information retrieval apparatus comprising:   The information search apparatus according to claim 1, wherein the keyword extraction unit and the index generation unit perform processing every predetermined period. The keyword history database further stores an address of a site where the keyword was posted and an extraction date and time of the keyword from the site,
When the keyword extraction unit writes the word, the keyword, and the weight in the keyword history database, an address of a site on which the keyword is posted and an extraction date and time of the keyword from the site Write further,
The information search apparatus according to claim 1, further comprising a keyword history database maintenance unit that deletes data that has passed a predetermined time from the keyword history database.   The information search apparatus according to claim 1, wherein the keyword history database maintenance unit performs processing every predetermined period. An information retrieval apparatus for reading a word corresponding to the received keyword from a word database in a storage area for storing a word and a keyword related to the word when the keyword is received from a client computer and transmitting the word to the client computer. On the computer,
A process of accessing a plurality of sites on the Internet designated in advance to obtain information on sentences posted on the sites;
A keyword related to a word stored in a keyword history database in a storage area for storing a word, a keyword related to the word, and a weight representing a high degree of association between the word and the keyword is extracted from the sentence information. Processing to
Calculating a weight that is a degree of association of the extracted keyword with the word, and writing the word, the extracted keyword, and the weight to the keyword history database;
Data is read from the keyword history database, the weight obtained by adding the word and the keyword together as a key to obtain the weight is obtained, and the word, the keyword, and the collected weight are stored in the Internet in the storage area. Writing to the index database;
When a keyword is received from the client computer, a word corresponding to the received keyword is read from the word database, and a degree of association between the received keyword and the word corresponding to the received keyword is stored in the Internet index database. A program for executing a process of determining a word corresponding to the received keyword and transmitting the degree of association to the client computer.

Images