JPH0973460A - Document retrieval device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently and accurately retrieve a desired document by the document retrieval device. SOLUTION: This document retrieval device is a document retrieval device which retrieves a desired document from a plurality of documents stored in a file 2 and is equipped with an input part 1 to which words, phrases, part-of- speech information, etc., are inputted, a storage means which adds words and phrases to part-of-speech information and saves them in the file 2, and a CPU 3 which retrieves a word and a phrase matching a key word and part-of-speech information inputted from the input part 1 from the file 2 and displays a document including the word and phrase at a display part 4.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document search device for searching a document stored in a storage means such as a file.

[0002]

2. Description of the Related Art Conventionally, a document search apparatus for performing full-text search for a document stored in a file or the like based on a certain keyword is well known. However, it is possible to store many documents in the file today. Then, improvement of document retrieval efficiency is desired.

Since many documents are usually registered in a file in a conventional document retrieval device, only one keyword in the document retrieval device is input and the same characters and phrases as those keywords are entered in the file. It takes a lot of time to search and display the document including the.

By the way, in this document retrieval apparatus, there are many cases where the document having the same characters and phrases as the inputted keyword does not exist in the file, and the document retrieval result is 0. In addition, there are cases where there are too many documents having the same characters and phrases as the input keyword in the file, and the meaning of the document search is meaningless.

In this case, the spent document search time is wasted, and the user inputs again a keyword similar to the previously input keyword into the document search device,
A new document search will be executed.

However, this not only results in poor document retrieval efficiency, but also results in a lot of input operations, resulting in a stagnation of work.

[0007]

As described above, in the conventional document search apparatus, when a document is searched for only with a certain keyword, there are many documents that do not have the same characters and phrases as those keywords, and There was a problem that the document search efficiency was poor because the document search was redone because it was too often.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a document retrieval apparatus capable of efficiently and accurately retrieving a desired document.

[0009]

In order to achieve the above-mentioned object, a document search device according to the invention of claim 1 is a document search device for searching a desired document registered in a file by a character string. Input means for inputting, morphological analysis means for morphologically analyzing the character string input from the input means into words and phrases, and part-of-speech information added to the words and phrases morphologically analyzed by the morphological analysis means Then, when the part-of-speech information is input together with the keyword from the input means during the document search, the registration means for registering in the file, and the corresponding words and phrases are searched from each document in the file based on these search conditions. And a display unit for displaying a document including the words and phrases searched by the search unit.

According to the first aspect of the invention, the words and phrases that have been morphologically analyzed by the morphological analysis means are:
Part of speech information is added to each and registered in the file.

When part-of-speech information is input together with a keyword from the input means at the time of document retrieval, the corresponding words and phrases are retrieved from each document in the file based on these retrieval conditions, and the document containing the words and phrases is retrieved. Is displayed. That is, it is possible to search for a document while limiting the search range by part of speech.

According to a second aspect of the present invention, there is provided a document retrieving apparatus for retrieving a desired document registered in a file by using a keyword, the input means inputting a character string and the input means. Morpheme analysis means for morpheme analysis of the character string into words and phrases,
A registration unit that adds part-of-speech information to each of the words and phrases that have been morpheme-analyzed by the morpheme-analyzing unit and registers it in the file; Retrieval means for retrieving corresponding words and phrases from each document of the file based on retrieval conditions, and counting means for counting the hit frequency of each word and phrase retrieved from the document by the retrieving means And a display unit for displaying a document including words and phrases whose frequency value counted by the counting unit is a predetermined value or more.

According to the second aspect of the invention, the words and phrases subjected to the morphological analysis are added with the POS information and registered in the file.

When part-of-speech information is input together with a keyword at the time of document search, the corresponding words and phrases are searched from each document in the file based on these search conditions. At this time, each time a word or phrase is searched from the document, the frequency of each hit is counted by the counting means, and a document including the word or phrase whose frequency value is a predetermined value or more is displayed. That is, the document can be searched while the search range is further limited by the part of speech and the appearance frequency.

As a result of the above, the desired document can be efficiently
And you can search accurately.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of the document retrieval apparatus of the present invention.

In FIG. 1, reference numeral 1 denotes an input unit such as a keyboard and a mouse, and various information such as symbols, numbers, characters (words such as katakana and hiragana) and character strings (words and phrases) defined by various ASCII codes and the like. , Keyword) and part-of-speech information (noun, verb, adjective, etc.) are input / instructed. Reference numeral 2 is a file, and the above various information is stored as a document after morphological analysis. 3 is CPU
The Japanese character analysis (including dictionary lookup / morphological analysis, syntactic analysis, etc.) is performed on the input character string similarity, and the various information described above is stored / registered as a document in the file 2 and input from the input unit 1. The corresponding document in the file 2 is searched based on the keyword information and displayed on the display unit 4.

Next, the operation of this document retrieval apparatus will be described with reference to FIGS. 2 and 4.

FIG. 2 is a diagram showing a word / phrase-part-of-speech correspondence table in the file 2, FIG. 3 is a flow chart showing a search operation of this document search device, and FIG. 4 is a diagram schematically showing how words are searched. . This document search device has a full-text search function. In this document retrieval apparatus, when the above-mentioned various information, for example, a character string is input by the input unit 1, the CPU 3 performs a morpheme analysis and divides the character string into meaningful characters and word units. Then, the part-of-speech information is added to each of the divided characters and phrases, and is stored / registered in the file 2.

At this time, the CPU 3 stores the file 2 in FIG.
As shown in, the table (index file) 20
Create This table 20 is a table in which part of speech information 22 is associated with each of the characters and phrases 21 subjected to morphological analysis.

For example, the words "document", "respect" and "respect" are associated with "nouns", the letters "respect" are associated with "verbs", and the phrase "relief" is associated with them. The "adjective" is associated.

In this document retrieval apparatus, when retrieving a document, the CPU 3 enters a document retrieval state by operating the input unit 1 in a predetermined manner, and the display unit 4 displays a retrieval character string (keyword).
Will be in the input standby state.

Here, it is assumed that the searcher thinks of a word he is thinking about, such as "respect", and a part of speech, such as "noun", in order to check whether the search keyword is appropriate. Then, for example, "respect" and part-of-speech information such as "noun" are input from the input unit 1 as keywords, and after the morphological analysis processing and the expression processing are performed by the CPU 3 (step 301 in FIG. 3). , The character string matching process (searching process) and the part-of-speech information matching process are executed for each document in the file 2 under the search condition in which "Kei" and "Noun" are combined (step 302-Step 303).

In this case, as shown in FIG.
In 0, there are three words and phrases that start with "respect", but two of them are "respect" and "respect" because the part of speech in search condition 41 is "noun". ,
A document containing these words is hit, and the hit document is stored in the output area of the CPU 3 (step 304).
). After that, the document is output from the CPU 3 to the display unit 4 and displayed.

As described above, according to the document search apparatus of this embodiment, when a document is registered, each word / phrase is added as part-of-speech information, for example, a noun, a verb, and an adjective, and is stored in the file 2. Sometimes, a user searches for a document by limiting the search range with keywords and parts of speech to search for a desired document.

That is, if the searcher inputs a word he / she thinks for checking whether or not the search keyword is appropriate, for example, "respect" and a part of speech, for example, "noun",
Among the plurality of documents registered in the file 2, since the documents that include the phrase that is a noun and the word that starts with "respect" are listed, the desired document can be searched efficiently and accurately.

Next, another embodiment will be described with reference to FIGS.

FIG. 5 is a diagram showing a relationship between a word / phrase registered in a file, a part of speech information, and the number of appearances, which is registered in a file by the document search apparatus of another embodiment, and FIG. 6 shows a document search operation of the document search apparatus of FIG. FIG. 7 is a flowchart schematically showing how words are searched for by this document search device.

In this embodiment, when the document is registered in the file 2 after the morphological analysis, the CPU 3 creates a table (index file) 50 as shown in FIG.

The table 50 is a table in which morphologically analyzed characters and phrases are associated with part-of-speech information, and the frequency information 51 is added each time these characters or phrases are hit during document retrieval. Has become.
The frequency information 51 is such that a numerical value is added each time a hit is made.

In this document retrieval apparatus, when a document is retrieved, the CPU 3 enters a document retrieval state by operating the input unit 1 in a predetermined manner, and the display unit 4 displays a retrieval character string (keyword).
Will be in the input standby state.

Here, as in the above-described embodiment, when a keyword such as "Kei" and part-of-speech information such as "Noun" are input from the input unit 1, the CPU 3 performs morphological analysis processing and expression processing. After being done (Step 60 in FIG. 6)
1) Then, under the search condition that combines "respect" and "noun", character string matching processing (search processing) and part-of-speech information matching processing are performed for each document in file 2 (step 602 to step 603), and frequency information matching processing is executed on the results obtained by these matching processing (step 604).

In this case, as shown in FIG.
In 0, there are three words that start with "respect". Of these, "nouns" as part-of-speech information are "respect" and "respect". Further, the frequency information of "respect" and "respect" is "16" and "6", respectively. In this case, the search condition 51 has a setting that the number of appearances is 10 or more, and a document that has 10 or more appearances from “respect” and “respect” is output, that is, a document that includes only one “respect” is output. Be subject to. Therefore, only the document including "respect" is stored in the output area of the CPU 3 (step 605). After that, the document is output from the CPU 3 to the display unit 4 and displayed.

As described above, according to the document search apparatus of this embodiment, when a document is registered, each word / phrase is added as part-of-speech information, for example, a noun, a verb, and an adjective, and stored in the file 2. Input unit 1 if desired by the user
If you enter the keyword and part-of-speech information from, the documents are searched for when the frequency (occurrence number) of these information hits is 10 times or more, so you can efficiently search for the desired document while limiting the search range. You can search accurately. Although the number of appearances is set to 10 or more in this document retrieval device, this value can be changed as desired at the time of initial setting.

The document search device is not limited to the above-mentioned full-text search. For example, a character string and a character string are sequentially input by the input unit 1 to create a document, store the document in the file 2 and register it, and then register an arbitrary character string for calling the document, that is, a keyword, which is a so-called keyword search. It can also be applied in cases.

In this case, a word having a meaning representative of the created document is generally registered as this keyword, but a word not existing in the document may be used. In this case, if this phrase is input from the input unit 2, the document can be directly searched.

[0038]

As described above, according to the document search apparatus of the first aspect, the part of speech information is added to the character and the phrase at the time of document registration and registered, and the search range is the keyword and the part of speech information at the time of document search. Since the document is searched in the limited range, the search range can be narrowed to search the desired document efficiently and accurately.

Further, according to the document search apparatus of the second aspect, when the document is registered, the part of speech information is added to the character and the phrase and registered, and at the time of the document search, the keyword, the part of speech, and the value of the frequency of hitting these pieces of information. Since a document is searched for in a more limited search range, a desired document can be searched efficiently and accurately by further narrowing the search range.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a relationship between a word / phrase registered in a file by this document search device and part-of-speech information.

FIG. 3 is a flowchart showing a document search operation of this document search device.

FIG. 4 is a diagram schematically showing how words are searched by this document search device.

FIG. 5 is a diagram showing a relationship between a word / phrase registered in a file by a document search device according to another embodiment, part-of-speech information, and the number of appearances (hit frequency).

6 is a flowchart showing a document search operation of the document search device of FIG.

FIG. 7 is a diagram schematically showing how words are searched for by this document search device.

[Explanation of Codes] 1 ... Input unit, 2 ... File, 3 ... CPU, 4 ... Display unit.

Claims (2)

[Claims] 1. A document retrieval device for retrieving a desired document registered in a file by a keyword, input means for inputting a character string, and morphological analysis of the character string input from the input means into words and phrases. A morphological analysis unit, a registration unit that adds part-of-speech information to each of the words and phrases that have been morphologically analyzed by the morpheme analysis unit, and registers it in the file; A search means for searching the corresponding words and phrases from each document of the file based on these search conditions, and a display means for displaying a document containing the words and phrases searched by the search means. A document retrieval device characterized by being provided. 2. A document retrieval apparatus for retrieving a desired document registered in a file by a keyword, input means for inputting a character string, and morphological analysis of the character string input from the input means into words and phrases. A morphological analysis unit, a registration unit that adds part-of-speech information to each of the words and phrases that have been morphologically analyzed by the morpheme analysis unit, and registers it in the file; When the search is performed, a search unit that searches corresponding documents and phrases from each document of the file based on these search conditions, and hits each time the search unit searches the document for the words and phrases. Counting means for counting the frequency, and the value of the frequency counted by the counting means is a word or phrase of a predetermined value or more And a display unit for displaying the included document.