JPH03123973A - Document retrieval method - Google Patents

Abstract

PURPOSE:To speed up the retrieval of a document even when a large number of the documents are retrieved by calculating the file likelihood of only the document containing a key word related to a retrieving key word by referring to an inverted file. CONSTITUTION:A key word connection table processing part 4 generates a key word connection table 8 in which relation between the related key word and the degree of association are described together with the necessary key word. All the key words related to the given retrieving key word are extracted by the table 8, and the list of the related key words is generated. The retrieving key word itself inputted by a user is included in this list. In respect of each element of the related key word list, the files containing these key words are checked from the inverted file 10, and the file likelihood of only the document related to these key words is calculated. The document number of the document whose likelihood has been calculated is written in a check list so as not to repeat the calculation of the same document plural times. Thus, only a portion proportional to the number of connections needs to be calculated.

Description

DETAILED DESCRIPTION OF THE INVENTION FIELD OF INDUSTRIAL APPLICATION The present invention is in the field of 11 document management, and particularly relates to a document search method for rapidly searching a large number of documents using keywords.

BACKGROUND OF THE INVENTION Conventionally, there have been various search methods for document retrieval devices, one of which uses a keyword connection table that describes information related to each keyword. First, this keyword connection table is used to search for related keywords from the search keyword given by the user. File certainty, which is an evaluation value for all documents, is calculated from these keywords (including search keywords). Therefore, the user's requested file accuracy is set as a threshold, and documents with file accuracy greater than or equal to the threshold are selected.
Alternatively, documents are selected in descending order of file certainty according to the number of documents requested by the user, and are used as the search results. In this way, documents with a high degree of relevance related to the search keyword are retrieved.

Problems to be Solved by the Invention However, according to such a processing method, it is necessary to calculate the file accuracy in proportion to the number of registered documents, resulting in a long processing time. Therefore, it is not suitable when a large number of documents are registered.

Means for Solving the Problem The system is equipped with a keyword connection table that is created when document information is registered in a file and describes related information between each keyword, and an inverted file that shows the relationship between registered documents and keywords. This inverted file is referred to, and the file accuracy is calculated only for documents containing keywords related to the search keyword, and documents are searched according to this file accuracy.

Since file accuracy is not calculated for documents that do not contain keywords related to the effect search keyword, the time required to calculate file accuracy is proportional to the number of connections for related keywords, not the number of registered documents. Even when searching for documents, the speed can be increased.

Example An embodiment of the present invention will be described based on the drawings.

FIG. 1 shows an example of the system configuration of a document retrieval device. An outline of its configuration and operation will be explained. First, when the keyword extraction unit 1 inputs the registered document 2, it extracts the keyword, and outputs the keyword and information about the registered document 2 to the document information management unit 3, the keyword connection table processing unit 4, and the inverted file creation unit 5. . The document information management unit 3 stores the extracted keywords and bibliographic information 6 in a file 7, and creates a database in a format that can be used during a search.

As shown in FIG. 2, the keyword connection table processing unit 4 creates a keyword connection table 8 that describes necessary keywords, the connections between related keywords, and the degree of association thereof, and stores it in a file 9. Here, the keyword connection table 8 has a list structure as shown in the example shown in FIG. 2, and is sorted in descending order of relevance.

The inverted file creation section 5 creates an inverted file 1O that associates each keyword with the registered document 2, as shown in FIG. 3, and stores it in the file 11. That is, the inverted file IO is a collection of pointers pointing from a keyword to a document containing that keyword.

Next, a document selection section 12 is provided. This document selection section 12 is intended to extract document files that are closer to the subject matter/required concept of the search user using a group of keywords. Output to. From among these, the user selects a keyword necessary for the requested subject, or selects a free keyword, and inputs a display search request again from the keyboard 13.

Such a document selection section 12 includes a request processing section 15, a sorting section 16, a display management section 17, and a keyword relationship speed calculation section 1.
8 and a file accuracy calculation section 19. Request processing unit 1
5 transfers the keyword received from the keyboard 13 to the keyword relation speed calculation section 18. The keyword relationship speed calculation unit 18 calculates the related keywords related to the transferred keywords and their related ↑ blue reports in the keyword connection table 8.
Extract from. The extracted keyword group is sorted by the sorting unit 16
The sorting unit 16 sorts the information in descending order of relevance and outputs it to the display management unit 17. The display management unit 17 outputs this related keyword group to the display 14 and displays it to the user. When the user further selects and inputs necessary keywords according to this display, the final keyword group is sent to the request processing unit 15 together with the document selection request.

Upon receiving the document selection request, the request processing section 15 causes the file probability calculation section 19 to transfer the keyword group. at the same time,
The keyword connection table processing unit 4 is instructed to change the weight of related information regarding the final keyword group. The file probability calculation section 19 calculates the file probability for the file 7 of the registered document 2 using the received keyword group, keyword connection table 8, and inverted file 10, and transfers the result to the sorting section 16.

When the calculation of the necessary file accuracy is completed, the sorting unit 16
The search results are sorted by , and then displayed on the display 14 as search results.

The object of the file accuracy calculation process in this embodiment will now be explained. First, all keywords that are related to a given search keyword are extracted using the keyword connection table 8, and a list of related keywords is created. This list also includes the search keywords entered by the user. For each element of the related keyword list, files that include these keywords are checked from the inverted file 10, and file certainty is calculated only for documents that are related. At this time, the document number for which the accuracy calculation was performed is written in the check list to avoid performing calculations multiple times on the same document. This allows calculations to be made in proportion to the number of connections for related keywords.

By the way, as the number of documents in the database increases, it is easily expected that the number of connections between keywords will also increase, and the method of this embodiment may also increase the number of documents to be calculated. In such a case, when extracting related keywords from the keyword connection table 8, only the top n keywords are extracted. This n indicates the search depth, and if the value of n is large, the file accuracy is calculated even for documents that have only keywords with a low search speed. Furthermore, instead of specifying the search depth n, only keywords having a search rate greater than a certain threshold may be retrieved. In any case, these operations are very easy because the keyword connection table 8 is sorted in descending order of connection speed. Furthermore, even if the relationship speed between keywords changes or new keywords are added, the list structure allows for easy changes.

Effects of the Invention As described above, the present invention calculates the file probability only for documents that include keywords related to the search keyword by referring to the Inva-Ted file. Since the file accuracy is not calculated for documents that are not included, the time required to calculate the file accuracy is proportional to the number of connections for related keywords, rather than the number of registered documents, making it easier to search for a large number of documents. Even if there is, the speed can be increased.

[Brief explanation of drawings]

The drawings show an embodiment of the present invention; FIG. 1 is a block diagram showing the system configuration of a document search device, FIG. 2 is an explanatory diagram showing an example of the contents of a keyword connection table, and FIG. 3 is an illustration of an inverted file. It is an explanatory diagram showing an example of contents. 8... Keyword connection table, lO... Inverted file

Claims (1)

[Claims] It is equipped with a keyword connection table that is created when document information is registered in a file and describes related information between each keyword, and an inverted file that shows the relationship between registered documents and keywords, and this inverted file is referenced. A document search method characterized in that the file probability is calculated only for documents containing keywords related to a search keyword, and documents are searched according to this file probability.