JP2003281182A - Information retrieval device, information retrieval method, program and recording medium - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To shorten the time required for extracting information fitting to a retrieval condition from information stored in a database. <P>SOLUTION: A data collection part 102 obtains related information corresponding to a data item specified by a setting file 200 from respective related information pieces of a document data stored in a groupware database 20a, and outputs the same to an indexing part 104. The indexing part 104 indexes the related information by each document data and registers the same in an index file stored in a database 10a for retrieval. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information search device, information search method, program and recording medium for searching information in a database.

[0002]

2. Description of the Related Art In companies, for example, LAN (Local Ar
ea Network) and other computer networks (hereinafter,
The network is simply referred to as "network"), and various data are shared in this network to improve work efficiency. Specifically, software called groupware or collaborative ware (hereinafter,
By introducing "groupware", connect to the network for various data (for example, shared documents and schedules of each user) held by this computer (hereinafter referred to as "groupware server"). It becomes accessible from each computer (hereinafter referred to as “client terminal”).

Further, the groupware has a function of searching the stored document data for the corresponding document data in response to a request from the client terminal. As a result, the user can easily find desired document data from the large amount of document data managed by the groupware server using the client terminal.

[0004]

However, when the groupware server searches for document data, it is common to execute a search process for all document data, and the number of document data and the number of document data There is a problem that the search time also increases in proportion to the capacity. In particular, in a call center that responds to inquiries from customers, this problem is exacerbated because the groupware server needs to quickly search and retrieve document data in response to inquiries from customers.

The present invention has been made in view of the above-mentioned circumstances, and an information search capable of shortening the time required to specify the information corresponding to the search condition among the information stored in the database. An object is to provide an apparatus, an information search method, a program, and a recording medium.

[0006]

In order to achieve the above object, the present invention associates at least text data containing a text sentence with identification information of the text data and also provides a plurality of associations related to the text sentence. In an information search device that searches a database that associates information, an item that classifies the plurality of related information items, and identification information of text data that corresponds to the text sentence, an item that can be a search target among the items is specified. First storing means for storing item specifying information to be stored, related information acquiring means for acquiring related information classified into items specified by the item specifying information from the database, and acquired by the related information acquiring means. A second storage unit that stores the related information in association with the identification information corresponding to the related information; A search condition acquisition unit that acquires a search condition conforming to the specified item by item specification information,
Provided is an information search device comprising: a search unit that specifies related information corresponding to the search condition from the related information stored in the second storage unit and specifies the identification information corresponding to the related information. To do.

In order to achieve the above object, the present invention associates at least text data including a text sentence with identification information of the text data, and
In the information search method for searching a database that associates a plurality of related information related to the text sentence, an item that classifies the plurality of related information, and identification information of the text data corresponding to the text sentence, among the items ,
A first step of storing, in a storage device, item designation information that designates an item that can be a search target; and a second step of acquiring, from the database, related information classified into the item designated by the item designation information. According to the third step of storing the related information acquired in the second step in the storage device in association with the identification information corresponding to the related information, and the item designated by the item designation information. The fourth step of obtaining the searched condition, and the related information corresponding to the searched condition is specified from the related information stored in the storage device, and the identification information corresponding to the related information is specified. An information search method including the process of 5 is provided.

According to the above-described information search device and information search method, only items that can be searched are extracted in advance from a plurality of items stored in the database, and the extracted items are searched. Done. Therefore, according to the present invention, the time required to identify the corresponding document data is shorter than when the search is performed for all the items in the database. Also, the user
The item to be searched can be changed only by changing the item designated by the item designation information.

Here, in the above information retrieval device, a body extracting means for extracting a text sentence from the text data, a morpheme analyzing means for dividing the extracted text sentence into a plurality of words, and each of the plurality of words. And an appearance frequency counting unit that counts the number of times that appears in the text sentence, and the second storage unit stores the word and the count value of the word as identification information of text data corresponding to the text sentence. It is desirable to store them in association with each other. According to this configuration, when a word is acquired as the search condition, the identification information of the text data can be specified in the order of including the word in large numbers.

In order to achieve the above object, the present invention associates at least text data including a text sentence with identification information of the text data, and
Among the above items, a computer that searches a database that associates a plurality of related information related to the text sentence, an item that classifies the plurality of related information items, and identification information of the text data corresponding to the text sentence is searched. First storage means for storing item designation information for designating an item that can be a target of, a related information acquisition means for obtaining related information classified into the item designated by the item designation information from the database, and the related information acquisition Second storage means for storing the related information acquired by the means in association with the identification information corresponding to the related information, and search condition acquisition for acquiring a search condition according to the item designated by the item designation information Means and related information stored in the second storage means, and specifies related information corresponding to the search condition. Provides a program for functioning as the search means for identifying the identification information corresponding to the related information. Of course, this program may be recorded in a computer-readable recording medium such as an optical disk or a magnetic disk.

[0011]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a diagram showing the configuration of an information search system according to an embodiment of the present invention. In this figure, the groupware server 20 is a groupware database 20a stored in a storage device such as a magnetic disk.
Is equipped with. This groupware database 20a
Document data shared by a large number of client terminals 30 connected via the network 2 is stored in the. Here, the document data is data including a text sentence. In addition, the groupware server 2
0 is a database in which shared document data is accumulated (that is, the groupware database 20 described above).
In addition to a), actually, various databases are provided, such as a database in which electronic mail data for each user is stored and a database in which schedule data for each user is stored.

Now, referring to FIG. 1, the information retrieval device 10
Consists of a personal computer, etc.,
A document data search request from the client terminal 30 is acquired via the network 2, and a candidate for the document data corresponding to this search request is transmitted to the client terminal 30. To explain further, the information retrieval device 10
Is equipped with a storage device such as a magnetic disk, and the search database 10a is stored in this storage device. The information search device 10 stores information related to each document data stored in the groupware database 20a in the search database 10a, and stores in the search database 10a when a search request is acquired from the client terminal 30. It is designed to search for information that has been posted.

FIG. 2 shows an information retrieval apparatus 1 according to this embodiment.
It is a functional block diagram which shows the structure of 0. In the figure,
The setting file analysis unit 100 specifies information to be accumulated in the search database 10a (hereinafter, referred to as “search information”) among the information related to the document data according to the instruction shown in the setting file 200, and collects the data. Part 1
Output to 02. Here, the setting file 200 is a data file created by, for example, the administrator of the groupware server 20, and the configuration thereof is shown in FIG.
As shown in the figure, each of the acquisition item, the weighted word, the storage destination address and the storage source address is designated in the setting file 200.

The acquisition item is for designating which of the data items managed by the groupware server 20 is to be acquired. More specifically, the groupware server 20 stores the groupware file 22 in which the related information related to the document data is recorded for each data item.
Prepared for each document data. FIG. 4 is a diagram showing an example of this groupware file. In this figure,
The character string “ITEM_NAME” indicates a data item, and the character string connected to the character string “ITEM_NAME” by an equal sign (=) indicates a data item name. For example, "ITEM_NAME =
In the case of “Classification”, the data item name is “Classification”. Further, the next line of the data item name (that is, the character string “ITEM_NAME”) is the related information of the document data corresponding to the data item name. In particular,
For example, the character string "TYPE_TEXT = Technical Note" written in the line following the character string "ITEM_NAME = Classification" is
Data item name “classification” Indicates that the related information of the corresponding document data is “technical note”. Therefore, the acquisition item specifies the data item name to be acquired, out of the data item names (data item name indicated by the character string “ITEM_NAME”) included in the groupware file 22. Although not shown, the groupware file 22 includes the groupware file 22.
It also shows which document data corresponds to.

The weighted words in the setting file 200 are for designating words that are frequently used as search words. The storage source address indicates the address where the database to be searched is stored. More specifically, the groupware server 20 generally includes a large number of databases as described above, and therefore, it is necessary to specify which database is to be searched. Therefore, the database to be searched is specified by designating the address. Further, the storage destination address indicates an address at which the information extracted according to the search information from each data in the database specified by the storage source address is stored in the search database 10a. In this way, by specifying different storage destination addresses for each database to be searched, it is possible to store the information extracted for each database to be searched in the multiple search database 10a. .

Now, referring to FIG. 2, the data collection unit 102
Receives the acquisition item indicated by the search information from the setting file analysis unit 100 from the groupware server 20 via the network 2 and performs the following processing. That is, the data collection unit 102 generates the body data file 202 from the items corresponding to the body part of the document data among the items acquired from the document data and the groupware file 22, and the information data from the items other than the body part. Generate file 204,
Each is output to the indexing unit 104. As shown in FIG. 5, in the body data file 202, a word (in the illustrated example, “interface device YYY” or the like) designated by the weighted word is added to the end of the body data (details will be described later). Further, as shown in FIG. 6, the information included in the information data file 204 includes, for example, a title (TITLE) attached to the document data and a storage source address (URL: Uniform Resource Locator) of the document data in the groupware database 20a. ) And so on. The function of the data collection unit 102 to acquire the document data from the groupware server 20 has an API (Application Program) provided by the groupware manufacturer.
m Interface).

The indexing unit 104 performs morphological analysis on the text data file 202 received from the data collection unit 102, and then performs indexing (indexing).
Is executed, and the execution result is the index file 20
The CPU in the computer to be registered in 6
Equivalent to. The index file 206 is stored in the search database 10a, and the index file 206 includes a page table 206a,
Keyword table 206c and word table 206
b is included (see FIG. 7). Each data table will be described later.

Here, the morphological analysis executed by the indexing unit 104 is to decompose a Japanese sentence written with kanji and kana mixed into words (morphemes) and specify the phonetic kana or part-of-speech of each word. is there. Morphological analysis dictionary 106
Is a dictionary used for morphological analysis in the indexing unit 104, and stores various words. More specifically, the indexing unit 104 decomposes the sentence into words (morphemes) by repeatedly extracting from the morphological analysis dictionary 106 the word that has the longest match with the subsequent portion of the sentence to be analyzed. Of course, if the text of the text data file is written in a language (for example, English) in which words are separated by spaces, morphological analysis is not necessary.

FIG. 8 is a diagram showing an example of the above-mentioned page table. The page table 206a is for managing information indicating the outline of each document data.
In one record of this page table 206a, document identification information, server identification information, storage source address,
The information includes last update date / time information, title information, text information, classification information, total word number information, software-specific document identification information, and reference level information.

Here, the document identification information is identification information uniquely assigned by the information search device 10 for each document data acquired from the groupware database 20a. The server identification information is information that identifies the groupware server 20 that is the acquisition source of the document data. In the present embodiment, as shown in FIG. 8, the information search device 10 uniquely allocates each server. Indicated by the number. The storage source address indicates the storage address of the document data in the groupware database 20a, and is specified by the URL as shown in FIG. The last update date / time information is information indicating the last date / time when the information search device 10 updated the information of the document data. The title information is information indicating the title (TITLE) of the document data, and is indicated by a character string having a predetermined number of bytes such as 256 bytes. The body information indicates a sentence of a predetermined number of characters (for example, 256 bytes) from the beginning of the body of the document data.

The classification information is information indicating the classification of the document of the document data. More specifically, for example, when the document data is shared by the network in the call center, the classification information indicates whether the document data is a product technical support document or a product manual. Information is recorded. The total word information indicates the total number of words in the text of the document data. The document identification information for each software is the groupware server 2
0 indicates unique identification information assigned to the document data. The reference level information indicates whether or not the browsing of the document data is limited to each client terminal connected to the network or is permitted to the terminals outside the network. Here, the server identification information and the software-specific document identification information are included in the page table 206a, because when the same groupware is installed in many servers, each server has the same identification information. Even when assigning to document data,
This is because it is possible to uniquely identify which document data of which server.

Next, FIG. 9 is a diagram showing an example of the above-mentioned word table. The word table 206b is for managing the words included in the body of each document data. More specifically, as shown in FIG. 9, in one record of the word table 206b, words, word identification information uniquely assigned to each word by the information search device 10, and accumulated in the groupware database 20a. Among all the document data, the word use document number indicating the number of document data including this word in the body is included. Here, the word usage document number is calculated according to the result of the morphological analysis of the text data file 202 of the document data by the indexing unit 104. Specifically, the indexing unit 104 performs a morphological analysis on one body data file 202 to convert the body into words (morphemes).
Then, each word is assigned unique identification information and registered in the word table 206b. And
The indexing unit 104 increments the value of the word usage document number corresponding to the registered word identification information by "1". As a result of such processing being performed on all the document data stored in the groupware database 20a, the number of word-used documents for each word can be obtained.

FIG. 10 is a diagram showing an example of the above-mentioned keyword table. This keyword table 2
06c is for each word included in the body of each document data,
This is for managing how many times a word appears. Specifically, as shown in FIG. 10, in one record of the keyword table 206c, the word identification information included in the word table 206b described above, the document identification information included in the page table 206a described above, and the number of appearances. And importance. The number of appearances indicates how many times the word appears in the text of the document data specified by the document identification information, and is obtained by the morphological analysis performed by the indexing unit 104. To further explain, the indexing unit 104 defines the body of the body data file 202 of the document data as a word (morpheme).
After being decomposed into, by counting how many words indicated by the word identification information are included in the body,
Calculate the appearance frequency. The degree of importance indicates the frequency of words in the body of all document data, and is calculated by the indexing unit 104 using the following formula. (Importance) = S × log (N / n) Here, S is the number of appearances, N is the number of document data accumulated in the groupware database 20a, and n is
This is the number of word-using documents described above. As shown in this equation, the more document data that includes the same word in the body, the less important the word becomes, and the more frequently the same word appears in the body of one document data, the more Words become more important. Here, as described above, since the weighting word is added to the end of the text data file 202 of the document data by the data collecting unit 102, the importance of the weighting word becomes relatively high. In particular, since the subject (TITLE) of the document data often includes words that remarkably reflect the content of the body of the document data, this subject is referred to as the body data file 20.
You may make it weight to 2.

In FIG. 2, the search request acquisition response unit 108
Receives a search request from the client terminal 30 via the network 2 and outputs the search request to the search unit 110. The search request acquisition response unit 108 corresponds to a network interface device in a computer. Also,
The search unit 110 searches the index file 206 stored in the search database 10a in response to the search request from the search request acquisition response unit 108, and returns the search result as
It outputs to the search request acquisition response unit 108. Upon receiving the search result from the search unit 110, the search request acquisition response unit 108 transmits the search result to the client terminal 30 via the network 2.

Next, the information search device 1 according to the present embodiment.
The operation of 0 will be described. Here, a program defining each processing procedure described below is stored in a recording medium such as a ROM or a magnetic disk included in the information search device 10. Note that this program may be recorded in a portable recording medium such as an optical disc, a magneto-optical disc, or a magnetic disc and installed in the information retrieval device 10, or the information may be transmitted via the network 2. It may be installed in the search device 10.

The information retrieval device 10 executes a registration process for registering in the index file 206 showing the information of each document data stored in the groupware database 20a. Specifically, as shown in FIG. 11, first, the setting file analysis unit 100 sets the setting file 2
00 is specified, the acquisition item, the weighted word, the storage source address and the storage destination address instructed by the setting file 200 are specified, and these specified information are output to the data collection unit 102 as search information (step S).
a1).

Next, the data collection unit 102 receives the acquisition item indicated by the search information from the setting file analysis unit 100 from the groupware server 20 via the network 2 and sends the text data file 202 (see FIG. 5) and The information data file 204 (see FIG. 6) is generated, and each is output to the indexing unit 104 (step Sa2).

The indexing unit 104 then receives the text data file 20 received from the data collection unit 102.
After performing morphological analysis on No. 2, indexing is executed, and the execution result is registered in the index file 206 including three data tables. (Step S
a3). As a result, the information about one document data is registered in the index file 206.
Next, the data collection unit 102 determines whether or not there is unprocessed document data in the groupware database 20a (step Sa4). If the determination result is YES, the information of the remaining document data is stored in the index file 206. The processing procedure is returned to step Sa2 so as to be registered in. On the other hand, if the determination result in step Sa4 is NO, the data collection unit 102 ends the process. As a result, the information of all the document data accumulated in the groupware database 20a is stored in the index file 2
It will be registered in 06.

By the way, the groupware database 2
The editing process is frequently performed such that the document data accumulated in 0a is added or deleted, or one document data is edited. Therefore, the information search device 10 uses the index file 20.
The following index file correction process is performed at regular time intervals so that the consistency between the information registered in 6 and each document data in the groupware database 20a is not broken.

That is, as shown in FIG. 12, first, the data collection unit 102 receives the acquisition item indicated by the search information from the setting file analysis unit 100 from the groupware server 20 via the network 2 and the body data. File 202 and information data file 204
Are generated and output to the indexing unit 104 (step Sb1). The indexing unit 104 uses the information registered in the text data file 202, the information data file 204, and the index file 206 to determine whether the document data is added, modified, or edited. It is determined whether or not it is not present (step Sb2).

More specifically, the indexing unit 104
Means that if the server identification information and the software-specific document identification information included in the information data file 204 are not registered in the page table 206a of the index file 206, this document data is added. Determine. On the other hand, the information data file 204
Although the information corresponding to the server identification information and the software-specific document identification information included in is already registered in the page table 206a of the index file 206,
If the last update date / time information differs between the information data file 204 and the index file 206, the indexing unit 104 determines that this document data has been modified. Furthermore, if each of the server identification information, the software-specific document identification information, and the last update date / time information is the same between the information data file 204 and the index file 206, the indexing unit 104 applies the document data to this document data. It is determined that no editing process has been performed.

Now, when it is determined that the determination result in step Sb2 is the added one,
The indexing unit 104 executes the same process as step Sa3 in the above-mentioned registration process, and registers the information of this document data in the index file 206 (step Sb3). Next, the data collection unit 102 determines whether there is unprocessed document data in the groupware database 20a (step Sb4). If the determination result is YES, the process is performed to process the remaining document data. The procedure is returned to step Sb1. As a result, the information on the document data added to the groupware database 20a is newly registered in the index file 206.

On the other hand, in the determination of step Sb2,
If it is determined that the document data has been corrected, the indexing unit 104 once deletes the information of the index file 206 corresponding to this document data, and then newly generates the information corresponding to this document data. Register in the index file 206. More specifically, the indexing unit 104 first identifies the document identification information (see FIG. 8) corresponding to this document data (step Sb).
5) Collectively delete the information about the specified document identification information from each of the page table 206a, word table 206b, and keyword table 206c included in the index file 206 (step Sb).
6). Next, the indexing unit 104 generates information corresponding to this document data by the indexing process described above and registers it in the index file 206 (step Sb7). Next, the data collection unit 102 determines whether there is unprocessed document data in the groupware database 20a (step Sb4), and if the determination result is YES, the process is performed to process the remaining document data. The procedure is returned to step Sb1. As a result, the correction made to the document data will be reflected in the index file 206. In addition, step Sb2
Even when it is determined that the edited result is that no editing has been performed, the indexing unit 104
Advances the processing step to step Sb4.

Next, if the decision result in the step Sb4 is NO, it means that the processing is executed for all the document data in the groupware database 20a. Therefore, during the series of processes described above, the document data corresponding to the document identification information that has never been referenced in the index file 206 (page table 206a) does not exist in the groupware database 20a. Therefore, the indexing unit 104 extracts all unreferenced document identification information from the page table 206a of the index file 206 (step Sb8), and the index file 206 includes each piece of information corresponding to the extracted document identification information. All the tables are deleted (step Sb9), and the process ends. As a result, the information corresponding to the document data deleted from the groupware database 20a is deleted from the index file 206. Further, when the document data is deleted, it is only necessary to delete the information corresponding to the document identification information from the index file 206, so that the time required to modify the index file 206 is shortened.

In this way, the index file 206
The information of each document data accumulated in the groupware database 20a is registered in the file, and even if the editing process such as addition, deletion, or correction is performed on the document data, the above-mentioned index file correction process is fixed. The information changed according to the editing process is immediately reflected in the index file 206 by being repeatedly performed every time.

When the search request acquisition response unit 108 of the information search apparatus 10 receives the search request from the client terminal 30 via the network 2, the search request acquisition response unit 108 outputs the search request to the search unit 110. The search unit 110 searches the index file 206 according to the received search request and extracts the information of the corresponding document data. More specifically,
In the search request, as a search word, a search word, or
The data items designated by the setting file 200 are included. For example, when a word is included in the search request as a search word, the search unit 110 refers to the keyword table 206c, and obtains the document identification information in the descending order of importance of the word (specifically, the word identification information). Extract. Then, the search unit 110 extracts, from the page table 206a, the title information, the body text information, the storage source address (URL), etc. corresponding to a predetermined number (for example, 20) of document identification information from the highest importance level. It is transmitted to the client terminal 30 via the request acquisition response unit 108.
As a result, the candidate of the document data corresponding to the search word is transmitted to the client terminal 30. Further, when the search request includes, for example, the last edit date and time as the search word, the search unit 110 causes the page table 206a.
Search each record of, and title information, body information and storage source address (URL) corresponding to the corresponding document identification information
Is transmitted to the client terminal 30 via the search request acquisition response unit 108. In the search request, as a search term,
Of course, each word and each data item may be included.

As described above, according to this embodiment, only the information that can be the search condition is registered in advance in the index file 206 for each document data stored in the groupware database 20a. Information retrieval device 1
0, when a search request is received, the index file 206 may be searched, and the data amount of the index file 206 is the groupware database 20.
Since the data amount is smaller than the amount of document data stored in a, the search can be performed faster than searching for each document data in the groupware database 20a. Further, if the user or the like changes the acquisition item designated by the setting file 200, the index file 206
Since the data items registered in can be changed,
The index file 206 can be configured according to the purpose of search. Further, the information search device 10 described in the present embodiment can be generally used among a plurality of groupware. More specifically, the index file 206 is constructed for each groupware simply by changing the acquisition item described in the setting file 200 for each groupware. Further, with this configuration, even if the setting file 200 is changed to construct the index file 206 for each groupware, the present embodiment is configured to operate the information search device 10 in association with the changed setting file 200. It is not necessary to recompile the program for information retrieval according to.

<Modification> The above-described embodiment is merely an example and shows one aspect of the present invention, and can be arbitrarily modified within the scope of the present invention. Therefore, various modifications will be described below.

For example, in the above-described embodiment, the configuration in which only one groupware server 20 is connected to the network 2 has been exemplified, but the configuration is not limited to this, and a configuration in which a plurality of groupware servers 20 are connected is also possible. good.
Furthermore, different groupware may be installed in each groupware server 20. More specifically, comprehensively searching each database of a plurality of different groupware servers generally requires different data management formats (for example, the number of data items and names) for each groupware. Have difficulty. On the other hand, in this modification, only the information of the data items that can be searched is registered in the page table 206a of the index file 206. Therefore,
Searching the page table 206a by the information search device 10 is equivalent to searching each database of a plurality of groupware servers.
A search of each database of a plurality of groupware servers is realized.

Further, for example, the indexing unit 104
When performing the morphological analysis on the body data file 202, the words that refer to the same thing such as “PC”, “personal computer”, and “personal computer” may be treated as one word. As a result, for example, even if "PC" is included in the search request as a search term, "PC" or "personal computer"
Document data including a word such as is also extracted as the corresponding document data, and the accuracy of search is improved.

[0042]

According to the present invention, an information search device capable of shortening the time required to specify the information corresponding to the search condition among the information stored in the database,
An information search method, a program, and a recording medium are provided.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an information search system according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a functional configuration of an information search device.

FIG. 3 is a diagram showing an example of the same setting file.

FIG. 4 is a diagram showing an example of the same groupware file.

FIG. 5 is a diagram showing an example of the text data file.

FIG. 6 is a diagram showing an example of the same information data file.

FIG. 7 is a conceptual diagram showing a data structure of the index file.

FIG. 8 is a diagram showing an example of the page table.

FIG. 9 is a diagram showing an example of the word table.

FIG. 10 is a diagram showing an example of the keyword table.

FIG. 11 is a flowchart showing a procedure of registration processing executed by the information search device.

FIG. 12 is a flowchart showing a procedure of an index file correction process executed by the information search device.

[Explanation of symbols]

10 ... Information search device, 10a ... Search database,
20 ... Groupware server, 20a ... Groupware database, 30 ... Client terminal, 100.
..Setting file analysis unit, 102 ... data collection unit, 10
4 ... Indexing unit, 106 ... Morphological analysis dictionary,
108 ... Search request acquisition response unit, 110 ... Search unit, 20
0 ... setting file, 206 ... index file.

Claims (5)

[Claims] 1. The text data including at least a text sentence and the identification information of the text data are associated with each other, a plurality of related information related to the text sentence, an item for classifying the plurality of related information, and the text. In an information search device for searching a database that associates identification information of text data corresponding to a sentence, a first storage unit that stores item designation information that designates an item that can be a search target among the items; Related information acquired by the related information acquisition unit for acquiring related information classified from the database specified by the specified information, the related information acquired by the related information acquisition unit,
A second storage unit for storing the identification information corresponding to the related information in association with the identification information; a search condition obtaining unit for obtaining a search condition according to the item designated by the item designation information; An information retrieving apparatus, comprising: retrieving means that identifies the relevant information corresponding to the search condition from the relevant information stored in the means, and identifies the identification information corresponding to the relevant information. 2. A body extracting unit that extracts a text sentence from the text data, a morpheme analyzing unit that divides the extracted text sentence into a plurality of words, and each of the plurality of words appears in the text sentence. An appearance frequency counting means for counting the number of times, wherein the second storage means stores the word and the count value of the word in association with identification information of text data corresponding to the text sentence. The information retrieval device according to claim 1, wherein the information retrieval device is a device. 3. The text data including at least a text sentence and the identification information of the text data are associated with each other, a plurality of related information related to the text sentence, an item for classifying the plurality of related information, and the text. In an information retrieval method for retrieving a database in which identification information of text data corresponding to a sentence is associated, a first step of storing item designation information for designating an item that can be a search target among the items in a storage device, A second step of acquiring, from the database, related information classified into items specified by the item specifying information; and the related information acquired in the second step, the identification information corresponding to the related information. According to the third process of associating and storing in the storage device, and the item designated by the item designation information. A fourth step of acquiring the search condition, and a fifth step of specifying related information corresponding to the search condition from the related information stored in the storage device and specifying the identification information corresponding to the related information. An information retrieval method comprising: 4. The text data including at least a text sentence and the identification information of the text data are associated with each other, a plurality of related information related to the text sentence, an item for classifying the plurality of related information, and the text. A computer that searches a database that associates identification information of text data corresponding to a sentence with a first storage unit that stores item designation information that designates an item that can be a search target among the items, according to the item designation information. Related information acquisition means for acquiring from the database the related information classified into the designated items, the related information acquired by the related information acquisition means,
A second storage unit that stores the identification information corresponding to the related information in association with the identification information; a search condition acquisition unit that acquires a search condition according to the item designated by the item designation information; and the second storage A program for causing related information corresponding to the search condition to be specified from among related information stored in the means, and causing the identification information corresponding to the related information to function as a searching means. 5. The text data including at least a text sentence and the identification information of the text data are associated with each other, a plurality of related information related to the text sentence, an item for classifying the plurality of related information, and the text. A computer that searches a database that associates identification information of text data corresponding to a sentence with a first storage unit that stores item designation information that designates an item that can be a search target among the items, according to the item designation information. Related information acquisition means for acquiring from the database the related information classified into the designated items, the related information acquired by the related information acquisition means,
A second storage unit that stores the identification information corresponding to the related information in association with the identification information; a search condition acquisition unit that acquires a search condition according to the item designated by the item designation information; and the second storage A readable-by-computer recording program for specifying the related information corresponding to the search condition from the related information stored in the means and causing the identification information corresponding to the related information to function as the search means. recoding media.