JP4980604B2 - Document search apparatus, document search method, document search program, and recording medium - Google Patents

Description

  The present invention relates to a document search device, a document search method, a document search program, and a recording medium, and more particularly to a document search device and a document search method for executing a search for a set of predetermined document information based on a search word and a related word. The present invention relates to a document search program and a recording medium.

  In the field of document retrieval, whether or not the retrieval result matches the retrieval request of the user is one of important evaluation criteria. Conventionally, related word expansion is known as a technique for obtaining a search result in accordance with a user's intention.

Related word expansion is a technique for adding to a search word a word that seems to be closely related to the search word. According to the related word expansion, it is possible to perform a search without omission as compared with a case where the search is performed only with a given search word. As a well-known related word expansion method, there is a method of searching for a synonym of a search word using a synonym dictionary and adding it to the search word (for example, Patent Document 1 and Patent Document 2). As another method, there is a method in which a search is first performed with a given search word, and a word included in the obtained search result is added as a co-occurrence word to the original search word (for example, Patent Document 3).
JP 2002-108914 A Japanese Patent No. 2850952 JP 2004-178421 A

  However, the former method has a limitation that only words in the synonym dictionary can be expanded. Therefore, there is a problem that complicated work such as maintenance of a synonym dictionary is required to deal with a new word. Further, the latter method has a problem that it can be determined as a co-occurrence word even though it is originally a word that hardly co-occurs when the search target data is as short as the title of a book. This is due to the fact that the search results contain fewer words and fewer co-occurrence words.

  The present invention has been made in view of the above points, and it is an object of the present invention to provide a document search device, a document search method, a document search program, and a recording medium that can output an appropriate search result in response to a search request. And

Accordingly, in order to solve the above-described problem, the present invention provides a document search apparatus that performs a search on a predetermined set of document information based on a search word and a related word, and inputs a plurality of search words, Input means for designating a second set of document information larger than the set of predetermined document information from the set of document information, and document information searched based on the search word from the set of second document information Related document search means for acquiring a set of related documents as related document information, and related word extraction means for extracting the related words from the related document information, the related document search means based on a predetermined attribute value, The number of document information for each attribute included in the set of document information searched based on the search word is calculated, and the document information searched based on the search word based on the calculated number of document information From the set of Select a set of document information belonging to the attribute, wherein the set of document information selected to the related document information.
Such a document search apparatus can output an appropriate search result in response to a search request.

  In order to solve the above problems, the present invention may be a document search method in the document search apparatus, a document search program for causing a computer to execute the document search method, or a recording medium on which the document search program is recorded. .

  According to the present invention, it is possible to provide a document search apparatus, a document search method, a document search program, and a recording medium that can output an appropriate search result in response to a search request.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram illustrating a hardware configuration example of a document search apparatus according to an embodiment of the present invention.

  1 includes a drive device 100, an auxiliary storage device 102, a memory device 103, an arithmetic processing device 104, a display device 105, an input device 106, and the like, which are mutually connected. Configured as follows.

  A program for realizing processing in the document search apparatus 10 is provided by a recording medium 101 such as a CD-ROM. When the recording medium 101 on which the program is recorded is set in the drive device 100, the program is installed from the recording medium 101 to the auxiliary storage device 102 via the drive device 100. The auxiliary storage device 102 stores the installed program and various data necessary for processing the program.

  The memory device 103 reads the program from the auxiliary storage device 102 and stores it when there is an instruction to start the program. The arithmetic processing device 104 executes functions related to the document search device 10 in accordance with a program stored in the memory device 103. The display device 105 displays a GUI (Graphical User Interface) or the like by a program. The input device 106 includes a keyboard and a mouse, and is used to input various operation instructions.

  Note that the document search apparatus 10 may be configured by a plurality of computers. For example, when the document search device 10 is configured as a client / server system, the display device 105 and the input device 106 are not necessarily required on the server side.

  FIG. 2 is a diagram illustrating a functional configuration example of the document search apparatus according to the first embodiment. In the first embodiment, an example in which similar searches are performed on the titles of books will be described.

  In FIG. 2, the document search apparatus 10 includes an input means 11, a related document search means 12, a related word selection means 13, a search means 14, an output means 15, a related word expansion database 16, a title search database 17, and the like. Is done.

  The input means 11 is a user interface for giving a search term to the system. It may be a keyboard, a display touch panel, a microphone and a voice recognition system. In other words, the user inputs, via the input means 11, a character string that seems to have some relationship with the title of the book to be searched for as a search term. This relationship means that the title contains the character string as it is, the title contains an expression similar to the character string, or the title contains a character string that can be used at the same time as the character string. It is to be.

  The related document search means 12 receives the search word from the input means 11 and searches the document information from the related word expansion database 16 based on the search word. A search is a word-by-word search by analyzing a search word into words, or a continuous length N character string that appears in the character string called an N-gram is obtained and the character string unit of that length N is obtained. Or search. The search result (hereinafter referred to as “related document information”) by the related document search means 12 is a title including a part of these search units, a title including all of them, or a score for the search word in each title. It may also be a title included in the top M.

  The related word expansion database 16 is a set of document information different from the title search database 17. The related word expansion database 16 holds at least a list of document information (here, titles of books) to be searched. Index information indicating what number a search unit such as a word is included in a list of titles may be held. This index information enables a high-speed search.

  The related word selection unit 13 extracts a character string related to the search word given by the user from the words constituting the related document information as the related word, and adds the interlaced word to the search word. The criteria for selecting the related word are, for example, selecting one having a high appearance frequency in the obtained related document information (for example, a set of titles), and the appearance frequency in a set of titles not included in the related document information. It may be such that a lower one is selected.

  The search means 14 searches the title of the book from the title search database 17 based on the search word to which the related word is added.

  The title search database 17 is a database in which a set of document information to be searched is stored. Therefore, in the present embodiment, the title search database 17 holds at least a list of titles. Also, index information may be held. The contents of the stored title and the type of information other than the title may be different from the related word expansion database 16, but it is desirable that the number of records in the related word expansion database 16 is larger ( That is, it is desirable that the set of document information stored in the related word expansion database 16 is larger than the set of document information stored in the title search database 17).

  The output means 15 displays the search result by the search means 14 on the display device 105, outputs it to a printer, or outputs it to a speaker as sound. Note that the finally obtained search result (search result by the search means 14) is a list of titles. Since the related words are expanded, the search word input by the user does not necessarily have to be included in the title.

  Hereinafter, the processing procedure of the document search apparatus 10 in FIG. 2 will be described. FIG. 3 is a flowchart for explaining search processing by the document search apparatus according to the first embodiment.

  In step S101, the input unit 11 receives an input of a search term from the user. Here, it is assumed that the user is looking for a book related to “patent”. Therefore, the user inputs the search term “patent” using the input means 11.

  In step S102 following step S101, the related document search unit 12 searches the related word expansion database 16 for related word expansion document information (related document information) based on the search word input by the user. . Specifically, the related document search means 12 receives the search word “patent” from the input means 11 and analyzes the search word in search units (in this example, the index of the related word expansion database 16 is configured in word units). Therefore, the search unit is a word.) As a result, the related document search means 12 obtains an analysis result that the word “patent” is included in the search word “patent” (in this example, the word set obtained by analyzing the search word is “patent”. A word set can be obtained if the search word consists of multiple words. Subsequently, the related document search unit 12 acquires a set of titles including the word “patent” from the related word expansion database 16 as related document information. FIG. 4 is a diagram illustrating an example of related document information as a search result by the related document search unit.

Progressing to step S103 following step S102, the related word selection means 13 extracts a related word from related document information. Specifically, for example, the following words are extracted from a set of titles as shown in FIG.
Standard, text, property, design, trademark, industry, possession, ...
The related word selection unit 13 adds the extracted related words to the search word input by the user (S104).

  Following step S104, the process proceeds to step S105, where the search unit 14 uses the search term to which the related term is added (that is, a set of words including the search term input by the user and the related term) for title search. The title of the book is searched from the database 17 (in this example, the index of the title search database 17 is also configured in units of words. Therefore, it is not necessary to re-analyze the search word).

  FIG. 5 is a diagram illustrating an example of a search result of book titles. As shown in the search result of FIG. 5, the search is executed using not only the search term (“patent”) but also the related term. Therefore, even if the title does not include “patent”, the relationship with “patent” A set of titles determined to be strong can be obtained.

  Progressing to step S106 following step S105, the output unit 15 causes the display device 105 to display a list of acquired titles.

  As described above, according to the document search apparatus 10 in the first embodiment, related words for extracting related words are extracted from a database different from the search database (title search database 17). Therefore, for example, if the database used for related word expansion is larger than the database for search, the number of co-occurrence words for the search word input by the user can be increased. Can be selected. Therefore, it is possible to reduce the possibility that an accidentally co-occurring word is selected as a related word and an unnecessary related word is added to the search word.

  In addition, since a synonym dictionary or the like is not required, it is possible to suppress the occurrence of costly work such as maintenance of the synonym dictionary.

  In the first embodiment, a plurality of related word expansion databases may be prepared, and the document search apparatus 10 may select a database suitable for the related words. For example, if the input search word is katakana, a computer-related or chemistry-related book database may be selected, and if it is an English character, a Western book database may be selected. As a result, the appearance probability itself of words that do not co-occur can be kept low.

  Next, a second embodiment will be described. Also in the second embodiment, an example in which similar searches are performed on the titles of books will be described.

  FIG. 6 is a diagram illustrating a functional configuration example of the document search apparatus according to the second embodiment. In FIG. 6, the same parts as those in FIG.

  As shown in FIG. 6, in the second embodiment, a plurality of databases (related word expansion databases 16a, 16b, 16c, etc., hereinafter collectively referred to as “related word expansion databases”) are used as related word expansion databases. Database 16 ”). Each of them includes at least title information of the book. Moreover, the input means 11 in 2nd embodiment receives designation | designated of the database used as the acquisition destination of a related word other than a search word from a user.

  Hereinafter, the processing procedure of the document search apparatus 10 in FIG. 6 will be described. FIG. 7 is a flowchart for explaining search processing by the document search apparatus according to the second embodiment.

  In step S201, the input unit 11 receives an input of a search word and identification information of a database from which a related word is acquired from a user. That is, the user designates which of the related word expansion databases 16a, 16b, and 16c is used as a related word acquisition source.

  Progressing to step S202 following step S201, the related document search means 12 searches related document information from the related word expansion database 16 designated by the user based on the search word.

  Since step S203 and subsequent steps are the same processing as step S103 and subsequent steps in the first embodiment, the description thereof is omitted.

  As described above, according to the document search device 10 in the second embodiment, a database suitable for the expansion of related words is selected by the user's judgment, so the appearance probability itself of words that do not co-occur in nature. Can be kept lower.

  Note that information stored in a plurality of related word expansion databases 16 may be classified according to predetermined attributes (for example, the field of books) related to document information. In such a case, the field of the book and each database may be associated in advance, and the user may be allowed to specify the field of the book instead of the identification information of the database.

  Next, a third embodiment will be described. Also in the third embodiment, an example in which similar searches are performed on the titles of books will be described.

  The functional configuration example of the document search apparatus 10 in the third embodiment may be the same as that in the first embodiment (FIG. 2) or the second embodiment (FIG. 6). However, in the related word development database 16, in addition to the title of each book, information indicating the field of the book (hereinafter referred to as “field information”) is also stored.

  The processing procedure of the document search apparatus 10 in the third embodiment is basically the same as that in the first embodiment (FIG. 3) or the second embodiment. However, since the related word expansion database 16 includes field information, the related document information acquired by the related document search unit 12 in step S102 or S202 includes field information in addition to the title. Yes.

  Further, the related word extraction processing (S103 or S203) from the related document information including the field information is performed as follows.

  That is, the related word selection means 13 does not simply acquire related words from the top N titles obtained in the search for related word expansion, but uses the following method for target word expansion (related word Related document information to be extracted).

FIG. 8 is a diagram for explaining a method of selecting target data for related word expansion. As shown in FIG. 8, the target data for the related word expansion is selected by the following procedure.
(1) Acquire the top N titles and fields.
(2) For each field, calculate the number of cases belonging to the field. In the example of FIG. 8, “practical” is calculated as 10 cases, “learning” is calculated as 4 cases, and “entertainment” is calculated as 2 cases.
(3) A set of titles of books belonging to the field having the largest number of calculated cases is set as target data for related word expansion. In the example of FIG. 8, a set of titles of books belonging to “practical” is the target data for related word expansion.

  Further, instead of acquiring only from the largest number of fields, weighting according to the number of cases in each field may be performed to acquire related words. Alternatively, search results including authors and publishers may be acquired and classified based on authors and publishers instead of fields. In addition, a search result including a publication date may be acquired, and a recently published book may be given a high weight. Moreover, you may utilize combining these.

  Thereafter, related words are extracted from the set of titles selected in this way, and the same processing as that after step S104 (or S204) is executed.

As described above, according to the document search apparatus 10 in the third embodiment, data other than the title of the book field or author, etc. that does not become a related word is put in the database used in the related word expansion, and these are related. Used to determine words. Therefore, it is possible to further suppress the possibility that a word that does not co-occur is selected as a related word and an unnecessary related word is added to the search word. Next, a fourth embodiment will be described. Also in the fourth embodiment, an example in which similar searches are performed on the titles of books will be described. FIG. 9 is a diagram illustrating a functional configuration example of the document search apparatus according to the fourth embodiment. 9, parts that are the same as those shown in FIG. 2 are given the same reference numerals, and explanation thereof is omitted.

  As shown in FIG. 9, the input means 11 in the fourth embodiment accepts designation of a field of a book for developing related words in addition to a search word from a user. In addition, in the related word expansion database 16 according to the fourth embodiment, in addition to the title of each book, information (field information) indicating the field of the book is also stored.

  Hereinafter, the processing procedure of the document search apparatus 10 in FIG. 9 will be described. FIG. 10 is a flowchart for explaining search processing by the document search apparatus according to the fourth embodiment.

  In step S401, the input unit 11 receives input of a search term and a field for developing related terms from the user. Progressing to step S402 following step S401, the related document search means 12 searches related document information from the related word expansion database 16 designated by the user based on the search word. The related document information acquired here includes field information in addition to the title.

  Progressing to step S403 following step S402, the related word selection means 13 extracts a related word from the related document information. Here, the related word selection unit 13 first extracts a set of titles related to the field designated by the user from the related document information including the field of the book.

  FIG. 11 is a diagram illustrating an example of extracting a set of titles related to a field designated by the user. FIG. 11 shows a set (N) of titles extracted when “practical” is designated as the field for developing related words.

  Subsequently, the related word selection means 13 extracts related words from the set of titles extracted as shown in FIG.

  The subsequent processes (S404 to S406) are the same as those in the first embodiment (S104 to S106).

  In addition, as a determination condition for related word expansion, a field other than the field of books may be used as described in the third embodiment. Further, the document search device 10 may allow the user to interactively specify the related word determination condition when acquiring the related word.

  As described above, according to the document search apparatus 10 in the fourth embodiment, with the assistance of the user, it is possible to set search conditions suitable for related word expansion and conditions suitable for determination of related words. The appearance probability and selection probability of a word that does not occur can be kept low, and the possibility that an unnecessary related word is added to the search word can be further reduced.

  In the above description, different databases include cases where tables are different within the same database system, as well as when the database systems are different. Therefore, it is not limited to being constructed in a separate storage device in terms of hardware.

  The document search apparatus in the first to fourth embodiments can be applied not only to the book search system but also to other search systems in general. However, it is desirable that the search target is as short as the title of the book. Further, for example, it is more desirable that the search target data such as patent documents, papers, news, and the like are classified to some extent and that many words are commonly used in each classification.

  As mentioned above, although the Example of this invention was explained in full detail, this invention is not limited to the specific embodiment which concerns, In the range of the summary of this invention described in the claim, various deformation | transformation * It can be changed.

It is a figure which shows the hardware structural example of the document search apparatus in embodiment of this invention. It is a figure which shows the function structural example of the document search device in 1st embodiment. It is a flowchart for demonstrating the search process by the document search device in 1st embodiment. It is a figure which shows the example of the related document information as a search result by a related document search means. It is a figure which shows the example of the search result of the title of a book. It is a figure which shows the function structural example of the document search device in 2nd embodiment. It is a flowchart for demonstrating the search process by the document search device in 2nd embodiment. It is a figure for demonstrating the selection method of the target data of related word expansion | deployment. It is a figure which shows the function structural example of the document search device in 4th Embodiment. It is a flowchart for demonstrating the search process by the document search device in 4th embodiment. It is a figure which shows the example of extraction of the collection of the titles concerning the field | area designated by the user.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Document search device 11 Input means 12 Related document search means 13 Related word selection means 14 Search means 15 Output means 16, 16a, 16b, 16c Related word expansion database 17 Title search database 100 Drive device 101 Recording medium 102 Auxiliary storage device 103 Memory Device 104 Arithmetic Processing Device 105 Display Device 106 Input Device

Claims (10)

A document search device that performs a search on a predetermined set of document information based on a search word and a related word,
Input means for inputting the search term and designating a second set of document information larger than the set of predetermined document information from a set of a plurality of document information;
Related document search means for acquiring, as related document information, a set of document information searched based on the search term from the set of second document information;
Related word extracting means for extracting the related word from the related document information,
The related document search means calculates the number of document information for each attribute included in the set of document information searched based on the search word based on a predetermined attribute value, and sets the calculated number of document information. And selecting a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term, and using the selected set of document information as the related document information. Document retrieval device.   2. The document search apparatus according to claim 1, wherein the related document search means selects a set of document information belonging to an attribute having the largest number of calculated document information and uses the set as the related document information.   The related document search means selects a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term based on a weighting according to the calculated number of document information. The document search apparatus according to claim 1, wherein the related document information is used. A document search method in a document search apparatus for executing a search for a predetermined set of document information based on a search word and a related word,
To enter the search word, the input procedure for specifying a set of second document information has greater than said set of predetermined document information from a set of a plurality of document information,
A related document search procedure for acquiring, as related document information, a set of document information searched based on the search term from the set of second document information;
A related word extraction procedure for extracting the related word from the related document information,
The related document search procedure calculates the number of document information for each attribute included in a set of document information searched based on the search word based on a predetermined attribute value, and sets the calculated number of document information. And selecting a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term, and using the selected set of document information as the related document information. How to search for documents.   5. The document search method according to claim 4, wherein the related document search means selects a set of document information belonging to the attribute having the largest number of calculated document information and uses the set as the related document information.   The related document search procedure selects a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term based on weighting according to the calculated number of document information. 6. The document search method according to claim 4, wherein the related document information is used. A document search program for causing a computer to execute a search for a set of predetermined document information based on a search word and a related word,
To enter the search word, the input procedure for specifying a set of second document information has greater than said set of predetermined document information from a set of a plurality of document information,
A related document retrieval procedure to retrieve a set of document information is retrieved on the basis from the collection of pre-Symbol second document information to the search word as a related document information,
A related word extraction procedure for extracting the related word from the related document information,
The related document search procedure calculates the number of document information for each attribute included in a set of document information searched based on the search word based on a predetermined attribute value, and sets the calculated number of document information. And selecting a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term, and using the selected set of document information as the related document information. Document search program.   8. The document search program according to claim 7, wherein the related document search procedure selects a set of document information belonging to an attribute having the largest number of calculated document information and uses the set as the related document information.   The related document search procedure selects a set of document information belonging to a predetermined attribute from a set of document information searched based on the search term based on weighting according to the calculated number of document information. 9. The document search program according to claim 7, wherein the related document information is used.   10. A computer-readable recording medium on which the document search program according to claim 7 is recorded.

Images