JPH103481A - Document retrieval device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a document retrieval device improving possibility that a desired document can be obtained from the document of a retrieval object. SOLUTION: A morpheme analysis part 3 analyzes a moptheme on the document in a document storage part 2 by using a noun dictionary in a noun dictionary storage part 11 and extracts a keyword. Then, an index file is generated and is additionally stored in a index storage part 4. The extracted keyword is added to the noun dictionary in a registered noun dictionary storage part 12. When a sentence being the retrieval object is inputted from the input part 1, the morpheme analysis part 3 analyzes the morpheme by using the noun dictionary in the registered noun dictionary storage part 12 and extracts the keyword. Furthermore, the morpheme is analyzed by using the noun dictionary in the noun dictionary storage part 11 and extracts the keyword. A specification part 5 executes retrieval by using the index file in the index storage part 4 based on the extracted keyword, and outputs a retrieval result from an output part 6.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search apparatus using a keyword search method.

[0002]

2. Description of the Related Art A keyword search method is generally used as a search method for searching a huge amount of documents. In the keyword search method, words that can be keywords are extracted in advance from documents to be searched and registered in a file called an index file. In the index file, a keyword and information such as a document name in which the keyword appears and a position in the document are recorded as a pair. At the time of retrieval, a keyword that matches a word input to specify a desired document is searched for in the index, and the document name or position information in the document is obtained, so that the document can be retrieved at high speed.

As a method of extracting a keyword from a document, there are a method of extracting the keyword manually and a method of automatic extraction using a keyword extraction technique represented by morphological analysis. From the viewpoint of the time cost of keyword extraction, it can be said that the method based on automatic extraction is advantageous.

[0004] A problem with the keyword search method is that words to be input to specify a desired document must be exactly the same as the keywords registered in the index file, and it is easy to specify appropriate input words. It is not a point.

In order to solve this problem, for example, Japanese Patent Laid-Open Publication No. Hei 7-182370 allows a sentence as an input for designating a desired document. A desired document can be obtained by automatically extracting a keyword from the designated text and comparing the obtained keyword with a keyword registered in the index file. According to this method, even when the input for designating the desired document does not completely match the word in the index file, it becomes possible to obtain the desired document if the word included in the input sentence matches, The above problems can be solved.
For example, if "I want to know copiers released by Company A" is specified as an input, the keywords "Company A", "Release",
“Copier” is automatically extracted, and a search is made for documents that include both of those keywords.

However, in the keyword automatic extraction technique including morphological analysis, the extracted keyword depends on the context. Therefore, when keywords are extracted by morphological analysis from a sentence specifying a document as in the above-mentioned document,
There is a problem that keywords different from the searcher's intention are extracted, and search omissions increase. For example, it is assumed that a description “A company affiliated car” exists in the search target document, and “A company affiliate” and “train” are extracted and registered by keyword extraction when creating an index file. Here, it is assumed that the searcher wants to search this document and designates “company A series”. If a morphological analysis is performed as in the above-mentioned document, "Company A" and "Series" may be automatically extracted and searched. In this case, a desired document cannot be obtained.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and has as its object to provide a document search apparatus which increases the possibility of obtaining a desired document from documents to be searched. It is assumed that.

[0008]

According to a first aspect of the present invention, in a document search apparatus, a document storage unit for storing a plurality of documents and a document extracted in advance from each document stored in the document storage unit. An index storage unit that stores an index file that records a document name in which the keyword exists together with the keyword, an extraction unit that analyzes a sentence input to specify a search target and extracts the keyword, and an extraction unit that extracts the keyword. Specifying means for specifying a document to be searched by comparing the keyword with the keyword in the index file stored in the index storing means; and reading the document specified by the specifying means from the document storing means. Output means for outputting, wherein the extraction means includes an index stored in the index storage means. It is characterized in that extracting the keyword by performing analysis using the keyword information present in the file.

According to a second aspect of the present invention, in the document search device according to the first aspect, the extraction means further does not use keyword information existing in an index file stored in the index storage means. And extracting keywords from the sentence.

According to a third aspect of the present invention, there is provided the document search apparatus according to the first aspect, wherein the first and second vocabulary dictionaries are stored in the document storage means using the first vocabulary dictionaries. Further comprising: an index creating unit that analyzes a document to extract keywords to create the index file, and adds the extracted keywords to the second vocabulary dictionary, wherein the extracting unit includes the second vocabulary dictionary. The sentence is analyzed to extract keywords.

[0011]

FIG. 1 is a block diagram showing an embodiment of a document search apparatus according to the present invention. In the figure, 1 is an input unit,
2 is a document storage unit, 3 is a morphological analysis unit, 4 is an index storage unit, 5 is a specific unit, 6 is an output unit, 11 is a noun dictionary storage unit, and 12 is a registered noun dictionary storage unit. In this embodiment, words extracted as keywords and registered in the index file are used as nouns. However, adjectives, verbs, and the like can be easily targeted as keywords.

The input unit 1 has a user interface through which a searcher can input a sentence for designating a desired document. The sentence for specifying the input document is sent to the morphological analysis unit 3.

The document storage unit 2 stores a plurality of documents to be searched.

The morphological analysis unit 3 performs a morphological analysis process on the document stored in the document storage unit 2 and the text input to the input unit 1 to extract keywords. The morphological analysis unit 3 stores a noun dictionary storage unit 11 storing a noun dictionary used when analyzing a document stored in the document storage unit 2, and a keyword extracted from the document.
It has a registered noun dictionary storage unit 12 used for analyzing a sentence to be searched. In addition, the morphological analysis unit 3
It also includes various vocabulary dictionaries other than grammar dictionaries and noun dictionaries necessary for executing morphological analysis.

When analyzing a document stored in the document storage unit 2, the morphological analysis unit 3 uses the noun dictionary storage unit 11
This is done by using the noun dictionary inside. An index file is created by recording a keyword extracted by analyzing the document and a document name including the keyword, and the index file is additionally stored in the index storage unit 4. Further, the keyword registered in the index file is added to the dictionary stored in the registered noun dictionary storage unit 32.

When analyzing a sentence input from the input unit 1, first, a keyword is extracted by using the noun dictionary in the registered noun dictionary storage unit 12, and a keyword is extracted and sent to the specifying unit 5. Then, analysis is performed using the noun dictionary in the noun dictionary storage unit 11 to extract keywords and send them to the specifying unit 5.

The noun dictionary storage unit 11 stores a dictionary describing a noun vocabulary necessary for morphological analysis. The stored noun dictionary is referred to when the morphological analysis unit 3 analyzes a document stored in the document storage unit 2 and a sentence input from the input unit 1.

The registered noun dictionary storage unit 12 stores a dictionary describing keywords obtained as a result of analyzing the documents in the document storage unit 2 by the morphological analysis unit 3. This noun dictionary is
The text input from the input unit 1 is referred to when the morphological analysis unit 3 analyzes the text. The nouns existing in the dictionary match the keywords in the index file stored in the index storage unit 4.

The index storage unit 4 stores the morphological analysis unit 3
And an index file that records keywords obtained from the documents stored in the document storage unit 2 by the analysis processing and document names including the respective keywords.

The specifying unit 5 refers to the index file in the index storage unit 4 for a document including all keywords obtained from the text input to the input unit 1 by the analysis processing of the morphological analysis unit 3, It is specified from the storage unit 2.

The output unit 6 has a user interface for displaying a document specified by the specifying unit 5 to a searcher as a search result.

FIG. 2 is a flowchart showing an example of the operation of the morphological analyzer 3 in the embodiment of the document search device of the present invention. In step S21, a sentence to be analyzed is received. In step S22, if the sentence is a document in the document storage unit 2, first, in step S23, a morphological analysis is performed using the noun dictionary in the noun dictionary storage unit 11, and a keyword is input. Extract. Next, in S24, the obtained keyword is recorded together with the document name including the keyword as an index file, and additionally stored in the index storage unit 4. Further, in S25, the keyword registered in the index file is added to the noun dictionary stored in the registered noun dictionary storage unit 12. Therefore, the nouns present in the dictionary stored in the registered noun dictionary storage unit 32 match the keywords in the index file stored in the index storage unit 4.

If the sentence to be retrieved received in S21 is a sentence input to the input unit 1, first in S26, a morphological analysis is performed using the noun dictionary in the registered noun dictionary storage unit 12 to extract keywords. I do. And S2
In step 7, the search result is obtained by notifying the specifying unit 5 of the result. Further, in S28, a keyword is extracted by performing morphological analysis using the noun dictionary in the noun dictionary storage unit 11, and the result is notified to the specifying unit 5 in S29, thereby further obtaining a search result. The search result is output from the output unit 6.

As described above, by performing morphological analysis using the noun dictionary in the registered noun dictionary storage unit 12 and extracting keywords, it is possible to increase the hit rate of keyword search. After that, a morphological analysis is performed using the normal noun dictionary in the noun dictionary storage unit 11 to extract keywords, thereby enabling a search with less omission.

For example, "A company affiliated car" is included in the search target document.
It is assumed that “company A” and “train” are extracted by keyword extraction when creating an index file and registered in the index file. Here, the searcher wants to search this document, and the input unit 1
Suppose "company A series" is specified. The morphological analysis unit 3 performs a morphological analysis process using the noun dictionary in the registered noun dictionary storage unit 12. In the noun dictionary in the registered noun dictionary storage unit 12, "Company A" extracted from the document is registered. Therefore, the word "Company A" registered by analysis of the designated "Company A series" is registered. And the unregistered word “column”, and the keyword “company A” is obtained. This makes it possible to search for a desired document.

[0026]

As is apparent from the above description, according to the present invention, it is possible to extract a keyword from an input sentence by giving priority to a keyword actually registered in an index file. Therefore, there is an effect that it is possible to prevent a search omission due to a context dependency of morphological analysis at the time of keyword extraction, which is a drawback of a search by automatic keyword extraction, and realize a search with a high hit rate. Further, by extracting keywords from the input text using a normal dictionary, search omissions can be reduced.

[Brief description of the drawings]

FIG. 1 is a configuration diagram illustrating an embodiment of a document search device according to the present invention.

FIG. 2 is a flowchart illustrating an example of an operation of a morphological analysis unit 3 in the embodiment of the document search device of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Input part, 2 ... Document storage part, 3 ... Morphological analysis part, 4 ...
Index storage unit, 5 ... specific unit, 6 ... output unit, 11 ...
Noun dictionary storage unit, 12 ... registered noun dictionary storage unit

Claims (3)

[Claims] A document storage unit for storing a plurality of documents;
An index storage means for storing an index file in which the name of a document in which the keyword exists is recorded together with a keyword previously extracted from each document stored in the document storage means, and a sentence input to specify a search target is analyzed Extracting means for extracting a keyword by extracting the keyword extracted by the extracting means and a keyword in an index file stored in the index storing means to specify a document to be searched;
Output means for reading and outputting the document specified by the specifying means from the document storage means, wherein the extraction means uses keyword information present in an index file stored in the index storage means. A document retrieval apparatus characterized in that a keyword is extracted by performing analysis. 2. The method according to claim 1, wherein the extracting unit performs analysis without using keyword information existing in an index file stored in the index storing unit to extract a keyword from the text. Item 2. The document search device according to Item 1. 3. The first and second vocabulary dictionaries, and a document stored in the document storage means are analyzed using the first vocabulary dictionaries to extract keywords and create and extract the index file. The second keyword is
2. The document according to claim 1, further comprising: an index creating unit that adds the keyword to the vocabulary dictionary, wherein the extracting unit analyzes the text using the second vocabulary dictionary and extracts a keyword. 3. Search device.