JPH09259133A - Document retrieval device and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To show a key word to a user to acquire retrieval candidates and also an effective key word to select these candidates. SOLUTION: A candidate document retrieval part 4b retrieves a document data file that includes a designated key word out of the group of document data files to be retrieved. A word relation data generation part 4c takes the document units such as sentences, paragraphs, etc., that include the designated key word out of the retrieved document data file and also extracts the words via analysis of morphemes to produce a word relation data which include the extracted words arranged in every document and in order of time. Then, a word list of the word relation data is synthesized, and the screens are shown in order of document production time such as order of later document production time or order of time designated by users, etc.

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 and a search method, and more particularly to a document search device and a search method suitable for searching all documents for searching a corresponding document file from a huge number of document files.

[0002]

2. Description of the Related Art Conventionally, as means for searching a document,
Keyword search systems are well known. This keyword search system searches for a document including a keyword input by a user. In order to search for a target document in this system, the user needs to specify an appropriate keyword included in the target document.

In this case, especially when there are a large number of documents in the database, a large number of candidate documents can be obtained only by designating a single keyword. Therefore, the user additionally inputs keywords to select candidates. It is necessary to perform a narrowed search such as narrowing down.

As one using such a retrieval system,
An all document search system has been developed in which a document file required by a user is searched from among a huge number of document files. In this document search system, in order to narrow down the search candidates, it is necessary for the user to connect some words with “or”, “and”, or “not”, create a search expression, and search accordingly. .

As described above, in the conventional search system, it is necessary to input a keyword when searching for a target document, and especially when a plurality of keywords are input to perform a narrowed search, these are recalled. Doing so was a burden on the user.

Also, when creating a search expression,
Since the relationship between words is unknown, it is difficult to combine appropriate words, and it takes a lot of time and effort to retrieve the target document.

[0007]

As described above, in the conventional document retrieval system, it is necessary to input a keyword when retrieving a target document, and in particular, a plurality of keywords are input to narrow down. When searching,
It was difficult for the user to properly select these keywords. For this reason, there is a drawback that much time and effort is required until the target document is retrieved.

The present invention has been made in view of the above circumstances, and provides a document search device and a search method capable of presenting to a user a keyword for obtaining a search candidate and a keyword effective for narrowing down the candidate. The purpose is to provide.

[0009]

The document retrieval apparatus of the present invention is a document retrieval apparatus for retrieving a document file containing a specified keyword from a document file group to be retrieved. Document unit extracting means for extracting a document unit such as a sentence or paragraph including the specified keyword from the document file, and morphological analysis is executed for each of the extracted document units to document the words contained in the document unit. Word extraction means for extracting each unit, word relation data storage means for storing the extracted words side by side in document units as word relation data related to the specified keyword, and word relation of this word relation data storage means And a word-related data display means for displaying the data on the screen in the order of creation time of the document.

In this document retrieval apparatus, first, a document file containing a designated keyword is retrieved from a document file group to be retrieved. Then, a document unit such as a sentence or paragraph including the designated keyword is extracted from each of the searched document files, and the words included in the document are extracted by morphological analysis. The extracted words are arranged for each document to create word-related data, which is stored in the word-related data storage means. Then, the word relational data of the word relational data storage means are displayed on the screen in the order of the document preparation time such as the order of the newest document preparation time or the time order specified by the user, and the words included in the word relational data are searched. It is presented to the user as keywords for obtaining candidates and effective keywords for narrowing down the candidates. Therefore, the user
It is possible to easily know the word related to the specified keyword and the temporal change of the related word, and use it for searching the target file.

Further, the word relation data display means includes a plurality of document units whose document creation times are adjacent in the word relation data of the word relation data storage means so that a predetermined number of words are displayed on the screen. It is preferable that the words in between are combined and displayed on the screen. Thus, when the number of words included in one word relation data is small, a plurality of word relation data can be combined and displayed, so that an appropriate number of words can be displayed on the screen. .

Further, with the word designated in the word relation data displayed on the screen as a new designated keyword, the document unit extraction, word extraction, word relation data storage, and word relation data display are performed. By further providing a means for recursively executing, information navigation that traces a network between words can be realized, and the user can perform document search without inputting a new keyword.

Further, by further comprising means for searching a document file including all the words designated in the word-related data displayed on the screen from the document file group to be searched, the screen can be displayed. With, it becomes possible to select a narrowed keyword, and the user does not need to input a new keyword when narrowing down search candidates.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows the configuration of a document search device according to an embodiment of the present invention. This document search device is for searching documents using a computer, and is composed of an input device 1, a display device 2, an external storage device 3, a control device 4, a memory 5, a communication device 6, and a clock device 7. There is.

The input device 1 is a device for inputting a character code, a control command, position information, etc., and is composed of, for example, a keyboard, a mouse, and a device for controlling these. The display device 2 is a device for displaying a prompt message for inputting by the user, an input character string, or document data obtained after a search, and is stored in, for example, a VRAM and this VRAM. It consists of a controller for displaying bit information as a dot string on the screen, and a display.

The external storage device 3 is, for example, a hard disk device, and is a device for storing document data files, search index data, and the like. As shown in FIG. 2, each document data file is composed of text data, non-text data such as an image, and document auxiliary information such as creation date / time information. A document ID number is assigned to each document data file. It is assigned.

The search index data is search information assigned to each document data file, and includes the keywords included in the corresponding document data file. As shown in FIG. 3, the search index data is stored in association with the document ID number of the original document data file.

The control device 4 is composed of, for example, a CPU, is connected to each hardware device through a bus, and controls each device, data transfer control between devices, a document search processing program, and the like. Execute the program. The document search processing program describes a procedure for document search processing, which is a feature of the present invention. Here, a keyword for obtaining a search candidate and a keyword effective for narrowing down the candidate are presented to the user. It includes a support function to do.

The memory 5 constitutes the main memory of this computer, and as shown in the figure, is a program section for storing a program for the control device 4 to execute various controls and processes, and is necessary for processing. It consists of a buffer for storing various data.

The above-mentioned document search processing program is stored in the program section. This document processing program includes a keyword input unit 4a, a candidate document search unit 4b, a word relation data generation unit 4c, and a word relation data display unit.
The keyword input buffer 4r and the candidate document number storage buffer 4 are included in the buffer unit used for the document search process.
s, a word relation data storage buffer 4t, and a display word list storage buffer 4u.

The word relation data storage buffer 4t is an array type buffer capable of storing a plurality of data and stores the word relation data. This word-related data corresponds to a specified keyword. Document units such as sentences or paragraphs containing the specified keyword are extracted from each document data file, and the words contained therein are arranged for each document unit in order of creation time of the document file. It is generated by This word relation data is composed of a plurality of elements corresponding to a plurality of document units. The structure of the word relation data storage buffer 4t is shown in FIG.

As shown in FIG. 4, one of the elements is divided into a date / time information storage section, an original text storage section, and a word list storage section. The date and time information storage unit stores the creation date and time of the document data file for the corresponding document unit, the original text storage unit stores the text data itself in the corresponding document unit, and the word list storage unit. Words extracted from text data in document units are stored.

The display word list storage buffer 4u is a buffer for synthesizing and storing the contents of a plurality of word list storage units of the word relation data storage buffer 4t, and the word group stored here is a designated keyword. Is presented to the user as a word associated with. The display word list storage buffer 4u is also in an array format capable of storing a plurality of data, and its structure is as shown in FIG.

Next, referring to the flowchart of FIG.
A specific document search processing procedure of this embodiment will be described. First, the keyword input unit 4a is activated, and the keyword input by the user is input to the keyword input buffer 4r (step 5a). Next, the candidate document search unit 4b
Is started (step 5b), the search index data stored in the external storage device 3 is referred to, all file search is executed, and the I of the document including the specified keyword is searched.
The D numbers are sequentially stored in the candidate document number storage buffer 4s. That is, in this candidate document search process, first, the search index data including the keyword specified by the user is checked, and if there is the search index data including the specified keyword, the corresponding document ID number is used as the candidate document number storage buffer. It is held for 4s.

In the subsequent step 5c, the word relational data generator 4c is activated. In the word relation data generation unit 4c,
Of the text data of each document data file obtained in step 5b, the part including the keyword specified in step 5a is extracted, sorted in the order of creation time of the document file, and stored in the original text storage section of the word relation data storage buffer 4t. Stored in order. Here, either "sentence unit" or "paragraph unit" can be set as the unit for extraction. Further, text creation date / time information is stored in the date / time information storage section of the word relation data storage buffer 4t.

Subsequently, the word relation data storage buffer 4t
Morphological analysis is performed on each of the text data stored in the original text storage unit of the above, and among the words extracted here, those other than the keyword specified in step 5a are stored in the word relation data storage buffer 4t. It is stored in the word list storage section.

In the subsequent step 5d, the word relation data display section 4d is activated. Here, the content of each element of the word relation data storage buffer 4t is referred to in order, the data of each word list storage unit is synthesized, stored in the display word list storage buffer 4u, and displayed.

In synthesizing the data, the total number of words after synthesizing is controlled to be constant. That is, if the number of words included in one word list is a fixed number, each element of the display word list storage buffer 4u corresponds to one word list element of the word relation data storage buffer 4t on a one-to-one basis. 1 of the relation data storage buffer 4t
The word list of the elements is stored in one element of the word relation data storage buffer 4t as it is, but when the number of words included in one word list element of the word relation data storage buffer 4t is small, the word list element adjacent to the word list element Is merged with the word.

The process of creating the display word list is shown in FIGS.
Shown in FIG. 7 shows an example in which a total of 7 or 8 words are displayed, and three words (words 5, 6, 7) are stored in the word list i + 1.
The four words (words 1, 2, ...) stored in the word list i arranged one before the word list i + 1.
3 and 4) are combined and stored in a certain element j of the display word list storage buffer 4u.

In this case, as shown in FIG. 8, if the word list i and the word list i + 1 include the same word (word 5), only one word 5 is a display word. It is stored in a certain element j of the list storage buffer 4u.

The process of creating such a display word list is as follows.
The word list is temporally newest, and the result is stored in the temporally newest word list element for display.

When the creation of the display word list is completed in this way, the display is performed. Here, first, as an initial state, data display of the last element (temporarily newest display word list element) of the display word list storage buffer 4u is performed.

The state of the screen at this time is shown in FIG. As shown, each word in the word list is displayed centering on the keyword specified in step 5a. Further, as shown in the figure, a slide button for changing the time of the word list to be displayed is also displayed on this screen.

In FIG. 6, the next state is waiting for input (steps 5e to 5f). Here, for example, when the slide button as shown in FIG. 9 is moved by the mouse or the like (step 5g), the process returns to step 5d, and after the screen is erased, the position of the moved slide button is designated. The contents of the display word list buffer element corresponding to the time are displayed.

For example, when the left end of the slide button is designated, the contents of the first element of the display word list buffer are displayed, and thereafter, each time the slide button is moved to the right, the display word list buffer is displayed. The elements after the are displayed in order.

When the word displayed on the screen is designated by the mouse or the like in step 5e, the following different operations are performed according to the designated operation mode (navigation mode, original document search mode).

(1) In case of navigation mode Step 5 is performed with the designated word as a new keyword.
A series of processing from b is executed recursively. 10 and 11 show the states of the screen when a word is designated and a word related thereto is displayed.

FIG. 10 shows a case where the word 7 is specified from the word group related to the keyword,
When a word on the screen is specified, the outline of the ellipse surrounding it is relatively thick. In this case, the word 7 is used as a new keyword to recursively execute the search processing routine of FIG. 6, that is, the above-described candidate document search, word relation data generation, and display word list generation, and the display word list is stored. The display of the last element (temporarily newest display word list element) of the buffer 4u is performed as shown in FIG.

(2) In case of original document search mode If one or more words (here, three) are designated as shown in FIG. 12 among the words displayed on the screen in step 5d, the candidate document search unit By activating 4b, the original document data including all the designated words is searched, and the result is displayed on the screen.

As described above, in this embodiment, the candidate document search section 4b first searches the document data file group to be searched for the document data file containing the specified keyword. Then, the word relational data generation unit 4c extracts document units such as sentences or paragraphs containing the designated keyword from each retrieved document data file and performs word extraction by morphological analysis to extract the extracted words into documents. Word-related data arranged in chronological order for each unit is created. Then, a word list of word-related data is synthesized and displayed on a certain number of screens according to the order of document creation time such as the order of newest document creation time, or the order of time specified by the user, and the words included in it are displayed. Is presented to the user as a keyword for obtaining search candidates and an effective keyword for narrowing down the candidates.

Therefore, the user can easily know the word related to the designated keyword and the temporal change of the related word, and can use it for searching the target file.

[0042]

As described above, according to the present invention, the words extracted from each candidate document are individually displayed on the screen,
It is possible to present the user with a keyword for obtaining a search candidate and an effective keyword for narrowing down the candidate, which realizes an improvement in operability for document search and a reduction in the time until the target document is searched. be able to. In addition, the word specified in the word-related data displayed on the screen is used as a new specified keyword to retrieve document units, extract words, store word-related data, and display word-related data recursively. As a result, information navigation that follows a network between words can be realized, and the user can perform document search without inputting a new keyword.

[Brief description of drawings]

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

FIG. 2 is an exemplary view showing a management structure of a document data file used in a document search process according to the first embodiment.

FIG. 3 is a diagram showing a management structure of search index data used in a document search process in the embodiment.

FIG. 4 is an exemplary view showing a data storage structure of a word relation data storage buffer used in a document search process in the embodiment.

FIG. 5 is an exemplary view showing a data storage structure of a display word list storage buffer used in a document search process according to the first embodiment.

FIG. 6 is an exemplary flowchart illustrating a procedure of document search processing according to the first embodiment.

FIG. 7 is an exemplary view for explaining an example of a display word list generation process performed in the document search process according to the first embodiment.

FIG. 8 is a view for explaining another example of the display word list generation process performed in the document search process according to the first embodiment.

FIG. 9 is a view showing a screen display example of a display word list in the same embodiment.

FIG. 10 is a view showing how a word is specified on the display word list screen in the embodiment.

FIG. 11 is a view showing a navigation mode screen obtained by recursively performing a search process using a word designated on the display word list screen as a new keyword in the same embodiment.

FIG. 12 is a view showing a state in which a plurality of words are specified on the display word list screen in the same embodiment.

[Explanation of symbols]

1 ... Input device, 2 ... Display device, 3 ... External storage device, 4 ...
Control device, 5 ... Memory, 4a ... Keyword input section, 4b
... Candidate document search unit, 4c ... Word relation data generation unit, 4d
... word relation data display section, 4r ... keyword input buffer, 4s ... candidate document number storage buffer, 4t ... word relation data storage buffer, 4u ... display word list storage buffer.

Claims (9)

[Claims] 1. A document search device for searching a document file including a specified keyword from a document file group to be searched, wherein each document file obtained as the search result is:
Document unit extracting means for extracting a document unit such as a sentence or paragraph including the specified keyword, and a word for extracting a word contained in the document unit for each document unit by executing morphological analysis on each of the extracted document units Extraction means, word relation data storage means for storing these extracted words as word relation data relating to the specified keyword side by side for each document, and word relation data of the word relation data storage means for document creation time A document retrieval device, comprising: a word-related data display means for sequentially displaying on a screen. 2. The word relation data display means, in the word relation data of the word relation data storage means, a plurality of document units whose document creation times are adjacent to each other so that a predetermined number of words are displayed on the screen. 2. The document search device according to claim 1, wherein the words in between are combined and displayed on the screen. 3. A word specified in the word relation data displayed on the screen as a new specified keyword,
The document retrieval apparatus according to claim 1, further comprising: a unit that recursively executes extraction of the document unit, word extraction, storage of word relation data, and display of word relation data. 4. A means for searching a document file including all words specified in the word-relationship data displayed on the screen from a document file group to be searched, further comprising: The document search device according to claim 1. 5. A document search device for searching a document file containing a specified keyword from a document file group to be searched, wherein each of the document files obtained as the search result includes:
Document unit extracting means for extracting a document unit such as a sentence or paragraph including the specified keyword, and a word for extracting a word contained in the document unit for each document unit by executing morphological analysis on each of the extracted document units Extraction means, word relation data storage means for arranging these extracted words in order of document creation time for each document unit, and storing them as word relation data related to the specified keyword, and this word relation data storage means A document retrieval device, comprising means for extracting word-related data corresponding to a specified arbitrary document creation time from the word-related data and displaying it on a screen. 6. A document retrieval method for retrieving a document file containing a specified keyword from a document file group to be retrieved, wherein each document file obtained as the retrieval result is
Document units such as sentences or paragraphs that include the specified keyword are extracted, morphological analysis is performed on each of the extracted document units, and the words included in the document units are extracted for each document unit. Is arranged in the order of creation time for each document unit to generate word-related data related to the specified keyword, and the word-related data is displayed on the screen in the order of creation time of the document. 7. A predetermined number of words are displayed on the screen,
7. The document search method according to claim 6, wherein in the word-related data, words between a plurality of document units whose document creation times are adjacent to each other are combined and displayed on the screen. 8. The word specified in the word relation data displayed on the screen is used as a new designated keyword to extract the document unit, extract the word, store the word relation data, and extract the word relation data. 7. The document search method according to claim 6, wherein the display is performed recursively. 9. The document file including all the words designated in the word-related data displayed on the screen is searched from the document file group to be searched for. Document search method.