JPH0793345A - Document retrieval device - Google Patents

Abstract

PURPOSE:To provide the document retrieval device which makes a user know what retrieval word and retrieval expression should be inputted and also, know what synonyms a word that the user has inputted has. CONSTITUTION:An index word consisting of a character string including a retrieval character string inputted from an input part 11 is retrieved by using a dictionary for analysis and the retrieval result is displayed at an input part 11, and the retrieval characters are selected according to the retrieval result displayed at the input part, the selected retrieval characters are retrieved in documents stored in a document data storage part 17, and the retrieval result is displayed at a retrieval result display part 16.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document retrieval apparatus for retrieving a document desired by a user from a document database storing a plurality of documents.

[0002]

2. Description of the Related Art Recently, a document information retrieval apparatus, particularly a document retrieval apparatus based on a full text search system has been realized. The full-text search method has an advantage that all characters or words in the document are searched, so that the labor for assigning the keyword is not required and the user needs to be familiar with the keyword.

[0003]

However, in the full-text search method, it is necessary to input all necessary search words in order to prevent omission of search.
Generally, it is difficult for a user to predict what kind of word is used in a document that he or she wants to search, and thus it is difficult to input an appropriate search formula. In other words, in order to search without omission of search, it is necessary for the user to enter as many search words as he or she can think of, and even in this case, it is possible that words other than the words that the user had come up with are used. is there. On the other hand, the retrieved documents include many documents that the user does not need.

Since it is generally difficult for a user to input all necessary search terms, a synonym dictionary is prepared and a search is performed using the search terms and synonyms of the search terms. However, it is very difficult to create a synonym dictionary because the synonym information varies depending on the user, the type of document, and the field of expertise.

The present invention has been made in view of the above points, and its purpose is to make it possible for a user to know what kind of search word / search expression should be input, and moreover, to add a word input by the user. An object of the present invention is to provide a document search device capable of recognizing what synonyms there are.

[0006]

A document search device according to a first aspect of the present invention includes a dictionary storing at least entry words,
Display means, document storage means for storing a plurality of documents, input means for inputting a search character string, and entry words composed of a character string including the search character string input by the input means, in the dictionary. A dictionary search means for searching from the search means, a means for displaying the search results searched by the dictionary search means on the display means, a selecting means for selecting a search character from the search results displayed on the display means, and a selecting means And a search unit for searching the document stored in the document storage unit for the search character selected in step 1, and displaying the search result searched by the search unit on the display unit. And

According to another aspect of the present invention, there is provided a document search device, which includes a dictionary storing at least an entry word, a display unit, a document storage unit storing a plurality of documents, and an input unit for inputting a search character string. A dictionary search unit that searches the dictionary for an entry word that matches a part of the search character string input by the input unit; and a unit that displays the search result searched by the dictionary search unit on the display unit. Selecting means for selecting a search character from the search result displayed on the display means; searching means for searching the search character selected by the selecting means from the documents stored in the document storing means; Means for displaying the search results searched by the means on the display means.

According to a third aspect of the present invention, there is provided a document retrieval device, a dictionary for storing word information constituting a compound word, a display means, a document storage means for storing a plurality of documents, and an input for inputting a retrieval compound word. Means, a dictionary search means for searching a compound word composed of the same constituent words as the constituent word constituting the search compound word input by the input means from the dictionary, and a search result searched by the dictionary search means On the display means, a selection means for selecting a search character from the search results displayed on the display means, and a search character selected by the selection means for a document stored in the document storage means. It is characterized by comprising a search means for searching from the inside and a means for displaying the search result searched by the search means on the display means.

According to another aspect of the present invention, there is provided a document retrieving apparatus, a display means, a document storing means for storing a plurality of documents, an input means for inputting a retrieval character string, and a retrieval character string input by the input means. A first search means for searching a character string adjacent to the document stored in the document storage means, means for displaying the search result searched by the search means on the display means, and the display means. Selecting means for selecting a search character from the displayed search results; second searching means for searching the search character selected by the selecting means from the documents stored in the document storing means; and the second searching means. Means for displaying the search result searched by the search means on the display means.

The document searching apparatus according to claim 5 is characterized in that the first searching means according to claim 4 searches for a document selected by the user by the selecting means or discarded.

According to a sixth aspect of the present invention, there is provided the first retrieval means according to the sixth aspect, wherein the characters immediately before or after the character string input by the user from the input means are in the same character type range. It is characterized by displaying only different ones.

[0012]

With the present invention, a headword formed of a character string including the search character string input from the input means or a constituent word forming a headword that matches a part of the input search character string or an input search compound word The dictionary is searched for a compound word composed of the same constituent word, the search result is displayed on the display means, the search character is selected by the selecting means from the search result displayed on the display means, and the selecting means is selected. The search character selected in step 1 is searched from the document stored in the document storage means, and the search result is displayed on the display means.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A document retrieval apparatus according to a first embodiment of the present invention will be described below with reference to FIGS. Figure 1
FIG. 1 is a block diagram showing the overall configuration of this document search device.
In FIG. 1, reference numeral 1 is a central processing unit that performs control and processing based on the present invention. The central processing unit 1 includes a display controller 3 for controlling a storage unit 2 such as a semiconductor memory, a magnetic disk, an optical disk, and a display unit 4 such as a liquid crystal display or a plasma display for displaying search results and document contents. An input unit 6 such as a keyboard and a mouse for inputting a user's search command and an input controller 5 for controlling the input unit 6 are connected.

Next, the configuration of the present embodiment will be described with reference to FIG. In FIG. 2, reference numeral 11 is an input section for inputting a keyword or a natural language sentence or sentence for instructing a search, converting an input of a user into a search command by an input analysis section, and outputting the result to a search section 15, After inputting the input of the person from the input unit 11 and converting the input into a search command, the search command is output to the input unit, and the input analysis units 12 and 13 that perform the synonym candidate creation process are input from the input analysis unit 12. An analysis dictionary, which is referred to and includes a Japanese analysis dictionary and a synonym dictionary, 14 is a control unit that controls each unit centrally, 15 is a search command input from the input unit, and is stored in the document data storage unit 17. A search unit that refers to the existing document data, searches for a related document set, and outputs the search result to the search result display unit 16. The search result display unit 16 inputs the search result and displays the document of the search result. 17 is a sentence Data and a document data storage unit for storing the index table of the document ID and the word for identifying the document data.

The structure of the analysis dictionary 13 is shown in FIG. 6, the search processing of the search unit 15 is shown in FIG. 12, and the structure of the keyword index table of the document data storage unit 17 is shown in FIG. It shows in.

Next, the operation of the first embodiment of the present invention constructed as above will be described. First, when an input sentence is input from the input unit 11, the input sentence is input to the input analysis unit 12.
The input analysis unit 12 performs the morphological analysis of the input sentence, the process of creating a search expression, and the process of creating synonym candidates.

FIG. 3 shows an example of the input sentence, the morphological analysis result of the input sentence in the input analysis unit 12, and the created search formula. As a method of morphological analysis, a method described in "Japanese Information Processing" (supervised by Nagao Shin, Institute of Electronics and Communication Engineers) and the like may be used instead of the main object of the present invention. Input analysis unit 1
2 takes out consecutive parts of nouns from the morphological analysis result and connects them through “*” to create a search formula. That is, as shown in FIG. 4, the input sentence "about the software for natural language processing", the search expression, and the words in the search expression are displayed on the input unit 11. Then, when the user specifies a search word on the window of the input unit 11 and specifies [synonym candidate display] with a mouse or the like, the input unit becomes the input analysis unit 1.
Start up 2. For example, the user specifies "soft" in the search word, and the square symbol of the line in which "soft" is displayed is highlighted.

The process of creating synonym candidates in the input analysis unit 12 will be described below. First, the configuration of the analysis dictionary 13 used for morphological analysis and creation of synonym candidates will be described with reference to FIG. As shown in FIG. 6, the analysis dictionary 13 is composed of a headline and a set of left connection code, right connection code, and part of speech used for analysis.

FIG. 7 shows a search dictionary structure for searching the analysis dictionary. As shown in FIG. 7, the search dictionary has a storage area for the first character of the heading of the analysis dictionary and 2
It consists of the storage area for the characters after the first character. Each storage area is a sequence of character numbers and characters, a sequence of character numbers of characters that can be connected backward, a separator 0, and a sequence of heading dictionary information (left connection code, right connection code, part of speech) consisting of the sequence of characters up to that point. , Separator 0. The characters in the storage area are arranged in the JIS code order, for example. Here, for example, in "soft", the dictionary information of the position of "to", which is obtained by tracing the character numbers of the characters that can be connected backwards starting from "so", is included in the existence of the "soft" heading and its dictionary information. Correspond. Furthermore, the dictionary information at the position of "A" obtained by tracing the character numbers of the characters that can be connected backward from "T" corresponds to "software". In addition, "A"
The dictionary information of the characters of “machine” obtained by tracing the character numbers of the characters that can be connected backwards corresponds to “software crisis”.

FIG. 8 shows the flow of the synonym candidate creation process described above. First, the first character of the search word is searched for in the first character area. If there are matching characters, follow them as long as there is a character number of characters that can be connected further. Print out all the headings you get along the way. In this way, synonym candidates as shown in FIG. 5 are displayed on the input unit 11.

The user can correct, delete, or add the displayed synonym candidate display processing result. FIG. 9 shows an example of the display after the user selects "software" from the synonyms of "software" in FIG. When the user specifies [Select], the input analysis unit 12 stores the correction result in a synonym dictionary and a buffer (not shown), and returns control to the input unit 11. The input unit 11 displays as shown in FIG. 10 according to the contents of the buffer.

The process of searching the document data storage unit 17 will be described below. FIG. 11 shows an example of the keyword index table of the document data storage unit 17. In this embodiment, the keywords in the document are stored in the keyword index table in the TRIE structure format. In this TRIE structure, in order to reduce the storage capacity and the trouble of searching, the same character string of each keyword is shared and stored. The keyword index table stores characters forming each keyword and link information between the characters.

For example, in correspondence with the keyword "machine", the link information of the character "machine" has the link information "00935". This link information represents the address where the character "machine" is stored. Then, in the character "machine" stored at the address "09935", the document data including the keyword "machine" is "file4", and another "machine" is included as a partial character string of the preceding two characters. “01201” is stored as link information about the keyword “machine translation”. By following this link information, the document data with "machine translation" as a keyword becomes "file25" and "file".
21 "(" file4 "and the like represent the document file name in which the document data is stored).

Further, for keywords including the same character string in front such as the keywords "actual example" and "experimental", two addresses stored in the link information of the character "actual" are "01003" and "01004". , Respectively, represent the addresses at which “test” and “example” are stored.

"0" in the link information represents an address or a separator of document data. The first characters of all keywords are sorted and stored in a fixed continuous storage area in the order such as JIS code order.

The search unit 15 searches the document file stored in the document data storage unit 17 for each search word in the search formula created by the input analysis unit 12, and the document file including the search word. Find the set of. Further, a set of document files that matches the search expression is obtained by a logical set operation of the document file set.

The processing of the search unit 15 will be described below with reference to FIG. First, as initialization processing, variables i and N are set to 1 and the number of search words is set, respectively. Variable i
Represents an index indicating a target search term. Variable i
While is smaller than the variable N, the “first character storage area” is searched for the first character of the search word i, the block in which the character is stored is obtained, and the block is designated as block A. Since the characters are sorted and stored in the "first character storage area", the block in which the character is stored can be obtained by a binary search.

Next, 2 is stored in the variable k. This variable k
Represents the character position of interest in the search word i. While the value stored in the variable k is smaller than the character string length of the search word i, the character stored in each block is compared with each character of the search word to find the corresponding block. . Finally, if there is document data corresponding to the search word, the block in the keyword index table can be detected correspondingly. The document file name stored in the link information of the block is set in the document file set i.

By performing the above process for all the search words, the document file sets 1 to N are set with the document file sets corresponding to the respective search words. In the next step, the common part of all the document file sets is found by the set operation and is set as the final document file set.

The title of the document of the document file set obtained by the search unit 15 is displayed on the search result display unit 16 and the search process is terminated. The analysis dictionary 13 shown in FIG.
In the search dictionary structure only for searching, only the character number sequence of characters that can be connected backward is stored in, but it is also possible to store the character number sequence of characters that can be connected forward. That is, in the storage area for the first character in FIG.
Stores the sequence of character numbers that can be connected to the front. Further, in the same structure as in FIG. 7B, a storage area is provided which stores a sequence of character numbers of connectable characters in the front instead of a sequence of character numbers of connectable characters in the rear. Using this dictionary structure, in the same way as the contents of the processing shown in FIG.
It is possible to search for a headword that includes the search word other than the position of the head.

Although the storage area for the first character and the storage area for the second and subsequent characters are divided in FIG. 7, the same storage area can be used. Further, although the above-described search unit 15 searches using a keyword index table composed of a word and a document ID including the word, which is created by morphological analysis of the document in advance, the document is not morphologically analyzed in advance as follows. It may be a method. First, a search is performed using a keyword index table composed of character places and document IDs containing the characters. Next, morphological analysis is performed on the sentences in the retrieved document, it is checked whether or not the retrieval word is included as a word in the document, and the document containing them is also displayed. Further, the document may be directly processed without using the keyword index table.

In the above description, the method of displaying the headword in the morphological analysis dictionary partially including the search word at the time of search is described, but the user registers the word in the analysis dictionary. You may make it display at a time.

As described above, in the first embodiment, in the full-text search system for morphologically analyzing a document and a search request sentence, an input search word is stored in a morphological analysis dictionary containing a character string of the search word. By displaying the heading word and allowing the user to search by adding the displayed heading word to the search word, the user can reduce the burden of first entering the search word without omission, In addition, omission of search can be prevented. Further, by storing the added search word in the synonym dictionary, it can be used at a later search.

Next, a second embodiment of the present invention will be described with reference to FIGS. 13 to 20. In this second embodiment,
In a full-text search system that performs a syntactic and semantic analysis of a document and a search request sentence, for an input search request sentence, a heading in a syntactic and semantic analysis dictionary that includes the syntactic and semantic structure of the search request part in the syntactic and semantic structure. The words and their syntactic and semantic relationships are displayed.

The device configuration and functional configuration of the second embodiment are the same as those of the above-described first embodiment shown in FIGS. 1 and 2 except for the following points. In the first embodiment described above, the input analysis unit 12
Although the morphological analysis was performed on the search command input from the input unit 11 and the document in the document data storage unit 17, in the second embodiment, the search command, the document element analysis, and the syntactic and semantic analysis are performed.

That is, the input analysis unit 12 performs a morphological analysis of an input sentence, a syntactic and semantic analysis, a search expression creating process, and a synonym candidate creating process. FIG. 13 shows a format and an example of a dictionary used in the input analysis unit 12. As shown in FIG. 13, the analysis dictionary is composed of a headline, a left connection code, a right connection code, a part of speech, and a set of semantic structures. ing. This analysis dictionary is used for morphological analysis, syntactic meaning and synonym candidate creation.

FIG. 14 shows an example of an input sentence, a morphological analysis result, a syntactic analysis result, a semantic analysis result, an unnecessary expression rule application result, and a created search expression of the input sentence in the input analysis unit 12. The method of morphological analysis and syntactic and semantic analysis is not the main point of the present invention, but "Japanese information processing".
(Makoto Nagao supervision, Institute of Electronics and Communication Engineers) etc. may be used. In applying the unnecessary expression rule, the unnecessary expression rule as shown in FIG. 15 is compared with the semantic analysis result,
Delete the part that matches the unnecessary expression rule from the semantic analysis result.

Next, the operation of the second embodiment of the present invention constructed as above will be described. When the search request character is input from the input unit 11, the input sentence, the search expression, and the words in the search expression are displayed as shown in FIG. The user
When [Synonymous candidate display] is specified with the mouse, etc.
1 activates the input analysis unit 12.

The input analysis unit 12 carries out a process of creating synonym candidates and transfers the result to the input unit 11. The input unit 11 displays the transferred result in the input unit window. FIG. 17 shows an example of the display of the synonym candidate creation processing result. In FIG. 17, “document information search” is displayed as a synonym candidate for “document search”.

The synonym candidate creation process will be described below with reference to the flowchart of FIG. That is, the words A and B having a direct semantic relationship in the search formula are taken out, and the word A, the word B, and the semantic relationship C between them are present in one semantic structure in the dictionary. Check whether or not. In this way, it is checked whether all two words in the search expression and the semantic relationship between them exist in one semantic structure in the dictionary, and if all exist, the semantic structure is input. Hand over to the department. This process is performed for all semantic structures in the dictionary. The synonym candidates retrieved in this way are displayed.

Then, the user can correct, delete, and add the displayed synonym candidates. FIG. 19 shows an example of the display after the user newly selects the search formula from the display of FIG. In the figure, when the user specifies [select], the input analysis unit 12 stores the correction result in a synonym dictionary and a buffer (not shown), and returns control to the input unit 11. The input unit 11 is
It is displayed as shown in FIG.

Then, the above-mentioned search processing is performed in the search section 15, and the title of the document file set obtained in the search section 15 and the like are displayed in the search result display section 16. It should be noted that the search method based on the morphological analysis and the syntactic and semantic analysis may use the method described in Japanese Patent Application No. 5-12561, not the main subject of the present invention. Further, in Japanese Patent Application No. 5-12561, a document is searched in advance using a keyword index table created by morphological analysis and syntactic and semantic analysis, but a method that does not analyze in advance may be used as follows. First, regarding the entered search word,
Whether a sentence in the retrieved document is morpheme-analyzed and syntactic-semantic-analyzed by using a keyword index table consisting of characters and the ID of the document containing the character, and whether the word and the syntactic-semantic analysis result are included in the document Check if As a result, only the document including the analysis result is displayed. It is also possible to directly analyze the document without using the keyword index table.

In the above explanation, the method of displaying the index word in the syntactic / semantic analysis dictionary partially including the syntactic / semantic structure of the retrieval request sentence at the time of retrieval has been described. It may be displayed when the word is registered.

As described above, according to the second embodiment, in the full-text search system for performing the syntactic and semantic analysis of the document and the retrieval request sentence, the syntactic and semantic structure of the retrieval request part is inputted into the syntactic and meaning structure for the inputted retrieval request sentence. Since the headwords in the syntactic and semantic analysis dictionary included in the structure and the syntactic and semantic relations are displayed, the user can search by adding the displayed syntactic and semantic relations to the search expression. it can. Therefore, the user can reduce the burden of inputting the search request sentence and can prevent the omission of the search. In addition, by storing the added words and syntactic and semantic relation expressions in the synonym dictionary and the synonymous expression dictionary, they can be used at the time of later retrieval.

Next, a third embodiment of the present invention will be described with reference to FIGS. The device configuration, the functional configuration, and the processing of morphological analysis and search expression creation in the input analysis unit 12 of the third embodiment are the same as those of the first embodiment described above, and therefore the description thereof is omitted.

First, when a search request input sentence is input from the input unit 11, it is displayed on the input unit 11 as shown in FIG. Then, when the user specifies the search word on the input section window and specifies [Synonym candidate display] with a mouse or the like, the input section 11 activates the input analysis section 12. In FIG. 21, the user specifies "software" in the search word, and the square symbol of the line in which "software" is displayed is highlighted.

The input analysis unit 12 performs the process of creating synonym candidates for the highlighted search word as described later with reference to FIG. 23, and the result is transferred to the input unit 11. The input unit 11 displays the transferred result in the input unit window. FIG. 22 shows an example of the display of the synonym candidate creation processing result. In FIG. 22, “software” and “ware” are displayed as synonym candidates for “software”.

Hereinafter, the synonym candidate creation process in the input analysis unit 12 will be described with reference to the flowchart of FIG. The format and example of the contents of the analysis dictionary 13 are the same as those in FIG. 6, and the format in which the analysis dictionary 13 is stored is the same as that in FIG. 7 except for the following points. That is, in FIG. 7, the character numbers of connectable characters are stored backward, but in this embodiment, the character numbers of connectable characters are stored forward.

First, the variable n is set to 1 and the search term 1 is set.
Search for the first character from the first character in the heading of the analysis dictionary. Then, when there is a matching character, if the headline information is stored, it is displayed and output.

Then, as long as there are character numbers of connectable characters in the backward direction, the character numbers of the connectable characters are traced until the number of characters of the search word matches n, and all the synonym candidates obtained in the process are displayed. There is.

In this way, the user can correct, delete, and add the displayed synonym candidate display processing result. FIG. 24 shows an example of the display after the user selects “software” from the synonym candidates of “software” in FIG. In FIG. 24, when the user specifies [select], the input analysis unit 12 stores the correction result in a synonym dictionary and a buffer (not shown), and returns control to the input unit 11. The input unit 11 displays as shown in FIG. 25 according to the contents of the buffer.

The above-described search unit 15 searches using a keyword index table composed of words and a document ID including the words, which is created by morphologically analyzing the document in advance. You can choose not to. First, a keyword index table consisting of characters and the ID of the document containing the characters is used for a search, the sentences in the searched document are morphologically analyzed, and it is checked whether or not the search word is included in the document. Display only documents that contain. Further, the document may be directly processed without using the keyword index table.

In the above description, the method of displaying the headword in the morphological analysis dictionary that matches the word constituting the search word at the time of search has been described. However, when the user registers the word in the analysis dictionary, May be displayed in.

In the third embodiment, a method of displaying a headword in the morphological analysis dictionary that matches a part of the input search word, such as the search word "software" and the headword "soft", will be described. It was Further, when a part of the search word and a part of the heading word in the morphological analysis dictionary match, the heading word may be displayed.
For example, when a compound word "machine translation" is entered as a search word, "automatic translation", "simultaneous translation", and "translation book" that include the word "translation" that composes the compound word The headword may be displayed so that the user can select it.

As described above, according to the third embodiment, in the full-text search system for morphologically analyzing a document and a search request, a headword in the morphological analysis dictionary that matches a part of the input search word. Since it is displayed, the user can search by adding the displayed headword to the search word. Therefore, the user does not need to input the search word first without omission, and thus the burden of inputting the search expression can be reduced and the omission of the search can be prevented. In addition, by storing the added search word in the synonym dictionary, it can be used in the subsequent search.

Next, the fourth embodiment of the present invention will be described with reference to FIG.
It will be described with reference to FIGS. The device configuration, the functional configuration, and the morphological analysis, the syntactic and semantic analysis, and the processing for creating the search formula in the input analysis unit 12 of the fourth embodiment are the same as those in the second embodiment described above, and thus the description thereof will be omitted.

The input analysis unit 12 performs a morphological analysis of an input sentence, a syntactic and semantic analysis, a search expression creation process, and a synonym candidate creation process. The format of the dictionary used in the input analysis unit 12, the example of the analysis result of the input analysis unit, and the example of the unnecessary expression rule are the same as those in FIGS. 13, 14 and 15.

When a search request input sentence is input from the input unit 11, as shown in FIG. 26, the input sentence, the search formula, and the words in the search formula are displayed. The user clicks [Synonym candidate display]
If you specify with a mouse, etc., the input unit 11
Start up 2. The input analysis unit 12 performs a synonym candidate creation process and transfers the result to the input unit 11. Input section 1
1 displays the transferred result in the input section window.
FIG. 27 shows a display of the synonym candidate creation processing result, and FIG. 27 shows a state in which “document search” is displayed as a synonym candidate of “document information search”.

The synonym candidate creation process will be described below with reference to FIG. First, the word A is extracted from the search formula and matched with the semantic structure in the dictionary. When it exists in the semantic structure in the dictionary, the word B having a direct semantic relationship with the word A and the relevant semantic relationship C are extracted from the semantic structure in the dictionary. The word B and the search expression are collated, and it is checked that there are two or more semantic relationships between the word A and the word B, and that the semantic relationship C is present therein. For one semantic structure in the dictionary, it is checked whether all two words having a direct semantic relation contained therein and the semantic relation between them exist in the search expression. Then, when all are present, the semantic structure is passed to the input unit 11. This process is performed for all semantic structures in the dictionary.

The user can correct, delete, or add the displayed synonym candidate display processing result retrieved as described above. FIG. 29 shows an example of the display after the user selects the newly displayed search formula in FIG. 27 and selects the example of the subsequent display.

When the user specifies [select] in the display state of FIG. 29, the input analysis section 12 stores the correction result in the synonym dictionary and a buffer (not shown), and returns the control to the input section 11. . The input unit 11 displays as shown in FIG. 30 according to the contents of the buffer.

In Japanese Patent Application No. 5-12561,
A document is searched using a keyword index table created by morphological analysis and syntactic and semantic analysis in advance.
A method that does not analyze in advance may be used as follows. That is, first, the input search word is searched using a keyword index table composed of characters and the ID of the document containing the characters, and then the sentence in the searched document is subjected to morphological analysis and syntactic and semantic analysis. Check if the word and syntactic and semantic analysis results are included in the document. As a result, only the document including the analysis result is displayed. It is also possible to directly analyze the document without using the keyword index table.

In the above description, the method of displaying the index word in the syntactic / semantic analysis dictionary that matches a part of the syntactic / semantic structure of the retrieval request sentence at the time of retrieval has been described. May be displayed at the time of registration.

In the above description, as in the case of the “document information search” and the heading word “document search” of the input sentence, the heading word of the syntactic and semantic structure that matches a part of the syntactic and semantic structure of the input compound word. Described how to display. Furthermore, when a part of the input sentence and a part of the heading word in the syntactic and semantic analysis dictionary match, the heading word may be displayed. For example, when a compound word “machine translation” is input as a search word, “automatic translation” or “automatic translation” that includes the word “translation” that constitutes the syntactic meaning structure of the compound word in the syntactic meaning structure The user may select a heading word such as “simultaneous translation” and select it. Naturally, not only the words in the syntactic and semantic structure, but also (tools)
When the relationships such as and (target) also match, the candidates can be synonymous. As described above, according to the fourth embodiment, in the full-text search system that performs the syntactic and semantic analysis of the document and the search request sentence, the input search request sentence matches a part of the syntactic and semantic structure of the search request sentence. Since the headwords in the syntactic-semantic-analysis dictionary of the syntactic-semantic structure and the syntactic-semantic relationship are displayed, the user can search by adding the displayed syntactic-semantic relationship to the search expression. Therefore, the user can reduce the burden of inputting the search request sentence and can prevent the omission of the search. In addition, by storing the added word or syntactic / semantic relation expression in the synonym dictionary or the synonym expression dictionary, it can be used at a later search.

In the first to fourth embodiments described above, when a word or a syntactic meaning structure in an input sentence partially matches an entry word or its syntactic meaning structure in the analysis dictionary,
The headword was displayed as a synonym candidate. In the following fifth to seventh embodiments, synonymous candidates are extracted and displayed from the document in the document data storage unit 17 instead of the analysis dictionary 13.

In the above-described first to fourth embodiments, the analysis dictionary is searched before the document search process, but the analysis dictionary may be searched after the document search process, for example. , The document search process is performed using the character string entered by the user, the user can judge by looking at the result, and can search the analysis dictionary only when necessary. The search process may be redone. Further, it is possible to process the document search processing and the analysis dictionary search in parallel. That is, a document search process is performed using the character string input by the user, and a parallel computer or the like searches the analysis dictionaries in parallel. Therefore,
The user can immediately see the search of the analysis dictionary when necessary.

Next, FIG. 31 shows the fifth embodiment of the present invention.
It will be described with reference to FIGS. The device configuration, the functional configuration, and the processing of morphological analysis and search expression creation in the input analysis unit of the fifth embodiment are the same as those of the first embodiment, and therefore description thereof will be omitted. The process of creating synonym candidates is performed by the search unit 15.

The keyword index table of the document data storage unit 17 will be described with reference to FIG. As shown in FIG. 33, in the keyword index table, the word and the document ID including each word are
Both the word and the word adjacent to the word are stored in the document. For example, the word "host" means file1
3 and file 82, the word "computer" is used after "host". The search unit 15 passes the information of what kind of word follows the input word by pulling this keyword index table to the input unit 11.

When a search request input sentence is input from the input unit 11, the input sentence, the search expression, the search word in the search expression and the constituent words forming the search word are displayed as shown in FIG. . When the user designates a constituent word on the input section window and designates [display synonym candidate] with a mouse or the like, the input section 11 activates the input analysis section 12.

In the case of FIG. 31, the user designates "host" in the constituent word, and the square symbol of the line in which "host" is displayed is highlighted. And the search unit 15
Performs a synonym candidate creation process and a document search process.
The search unit 15 performs a synonym candidate creation process. That is, the search unit 15 obtains information about what kind of word follows the input word by drawing this keyword index table, and passes it to the input unit 11. As a result, the input unit 11 displays the transferred result in the input unit window. FIG. 32 shows an example of the display of the synonym candidate creation processing result. As the synonym candidates of “host computer”, “host computer”, “host satellite system”, “host processor”, and “host centralized system” are displayed. It shows the displayed status.

Then, the user can correct, delete, or add the synonym candidate creation processing result displayed as shown in FIG. An example of a display after the user selects "host computer" and "host processor" from the synonyms of "host computer" in FIG. 32 is shown in FIG.
4 shows. In FIG. 34, when the user specifies [select], the search unit stores the correction result in the synonym dictionary and a buffer (not shown), and returns control to the input unit 11. Input section 1
1 is displayed as shown in FIG. 35 according to the contents of the buffer.

In the above-mentioned embodiment, the compound word including the constituent words forming the search word is extracted from the document and displayed, but the sentence including the compound word may be displayed. Further, the search unit 15 searches using a keyword index table that is created in advance by performing morphological analysis on a document and that includes a word and a document ID that includes the word. Good. First, the input search word is searched using a keyword index table composed of characters and the ID of the document containing the characters, and then the sentence in the searched document is subjected to morphological analysis and synonymous with the above. Extract and display word candidates. After that, the user selects a synonym. Further, the search word and the selected synonym are searched using the keyword index table of characters. Next, the sentence in the retrieved document is morphologically analyzed, it is checked whether or not the retrieval word is included as a word in the document, and only the document including them is displayed. Note that it is also possible to directly process and search the document without using the keyword index table.

In the above description, a method of displaying synonym candidates at the time of retrieval has been described, but they may be displayed at the time when the user registers a word in the analysis dictionary. Further, it is possible to search for a dictionary heading that includes the input constituent word of the compound word by using the dictionary that stores the information of the constituent word of the compound word as follows. First, FIG. 44 shows a format and an example of the dictionary. FIG. 45 shows the flow of synonym candidate creation processing in the input analysis unit 12. As shown in FIG. 45, for the constituent words that make up the compound word specified by the user, search for a compound word in which the order of arrangement of the constituent words in the compound word and the notation of the constituent words are the same, It is passed to the input unit 11. It should be noted that the analysis dictionary may be stored so that the constituent words can be searched for each rank of the arrangement of the constituent words. Further, as in FIG. 7 and the like, a constituent word is used as a heading, and the number of constituent words connectable to the rear or the number of constituent words connectable to the front is stored so that a compound word can be searched from the constituent words. You may

Further, in the fifth embodiment, the word adjacent to the search word is searched and displayed for all the documents in the document data storage unit 17, but the document is searched by the input search word and searched. The word adjacent to the search word may be searched and displayed only in the document. Furthermore, for any number of documents that the user has determined to be appropriate from the retrieved documents, the words adjacent to the search term may be retrieved and displayed only from among them. In addition, for any number of documents that the user determined to be inappropriate from the retrieved documents,
It is also possible to search and display a word adjacent to the search word from among them so that the user can use it for searching with a not condition.

As described above, according to the fifth embodiment, when the input search word is a compound word, the words forming the compound word are different from any other words in the document to be searched. Since it is displayed whether or not they are adjacent to each other, the user can add a necessary compound word or word to the search word so that the user can search. Therefore, the user does not need to input the search word first without omission, and thus the burden of inputting the search expression can be reduced and the omission of the search can be prevented. In addition, by storing the added search word in the synonym dictionary, it can be used in the subsequent search.

Next, the sixth embodiment of the present invention will be described with reference to FIG.
It will be described with reference to FIGS. The device configuration and functional configuration of the sixth embodiment, and the morpheme analysis, syntactic and semantic analysis, and search expression creation processing in the input analysis unit 12 are the same as those in the second embodiment, so description thereof is omitted. Document data storage unit 1
7. The description of the document search process in the search unit 15, the search result display unit, and the like are the same as those in the second embodiment, and are omitted here. The process of creating synonym candidates is performed by the search unit 15.

FIG. 38 shows the semantic structure index in the document data storage unit 17. The document data storage unit 17 further stores connection information having the semantic structure shown in FIG. 39, the meaning structure numbers shown in FIG. 38 are used to store which meaning structure is connected to which meaning structure. For example, after the semantic structure of 101,
It is shown that the semantic structures No. 301 and No. 405 are connected. By the number of the semantic structure and the connection information of FIG. 39,
You can search for semantic structures that connect to any semantic structure.

When a search request input sentence is input from the input unit 11, as shown in FIG. 36, the input sentence, the search formula, and the words in the search formula are displayed. When the user designates “display of synonymous candidates” with a mouse or the like, the input unit 11 activates the input analysis unit 12. In FIG. 36, the user has designated “translation” in the search word, and the square symbol of the line in which “translation” is displayed is highlighted.

Then, the search section 15 carries out a synonym candidate creation process and a document search process. The search unit 15 performs a synonym candidate creation process and transfers the result to the input unit 11. The input unit 11 displays the transferred result in the input unit window. FIG. 37 shows an example of the display of the synonym candidate creation processing result, and "automatic translation" and "human translation" are displayed as synonym candidates for "machine translation".

The user can correct, delete, and add the synonym candidate creation processing result displayed on the input section. FIG. 40 shows a display example after the user selects “automatic translation” from the synonymous candidates of “machine translation” in FIG. 37. In FIG. 40, when the user specifies [select], the search unit 15 stores the correction result in a synonym dictionary and a buffer (not shown), and returns control to the input unit 11. The input unit 11 displays as shown in FIG. 41 according to the contents of the buffer.

In Japanese Patent Application No. 5-12561,
A document is searched using a keyword index table created by morphological analysis and syntactic and semantic analysis in advance.
A method that does not analyze in advance may be used as follows. First, regarding the input search word, I and the document I containing the character
A keyword index table composed of D is used for searching, and then the sentence in the searched document is subjected to morphological analysis and syntactic and semantic analysis, and synonymous candidates are extracted and displayed in the same manner as described above. After that, the user selects a synonym candidate, and searches for the synonym candidate selected as the search term using the character keyword index table. Next, the sentence in the retrieved document is subjected to morphological analysis and syntactic and semantic analysis to check whether the word and the syntactic and semantic analysis result are included in the document. As a result, only the document including the analysis result is displayed. Note that it is also possible to directly perform document analysis processing without using the keyword index table.

In the above description, the synonym candidates are extracted from the document and displayed, but a sentence containing the synonym candidates may be displayed. Furthermore, in the above explanation, synonym candidates are
Although the method of displaying at the time of search is described, it may be displayed at the time when the user registers a word in the analysis dictionary.

In this embodiment, it is displayed what kind of word in the document the part constituting the result of the syntactic and semantic analysis of the input sentence is adjacent to. It is also possible to display what kind of word the search word or the entire syntactic and semantic analysis result of the input sentence is adjacent to in the document, instead of the constituent part.

Further, it is possible to search for a word including a part of the input search expression by using a dictionary that stores the information of the syntactic and semantic analysis. FIG. 46 shows the input analysis unit 12 in this case.
The flow of the synonym candidate creation process of is shown.

In the sixth embodiment, the document data storage unit 1
Words adjacent to the search word were searched and displayed for all the documents of No. 7, but the documents were searched by the input search word, and the words adjacent to the search word were searched and displayed only from the searched documents. You may. Furthermore, for any number of documents that the user has determined to be appropriate from the retrieved documents, the words adjacent to the search term may be retrieved and displayed only from among them. In addition, for any number of documents that the user has determined to be inappropriate from the retrieved documents, the word adjacent to the search word is retrieved and displayed, and the user selects no.
It can also be used to search with a condition of t.

As described above, according to the sixth embodiment, when the input search word is a compound word, the words forming the compound word are different from any other words in the document to be searched. Since it is displayed whether or not they are adjacent to each other, the user can add a necessary compound word or word to the search word so that the user can search. Therefore, the user does not need to input the search word first without omission, and thus the burden of inputting the search expression can be reduced and the omission of the search can be prevented. In addition, by storing the added search word in the synonym dictionary, it can be used in the subsequent search.

Next, the seventh embodiment of the present invention will be described with reference to FIG.
And FIG. 43. In the fifth embodiment described above, it is displayed what words in the document are adjacent to the words constituting the search word. In the seventh embodiment, however, what kind of character string matches the search word is displayed. It is displayed whether it is adjacent to the character string.

The device configuration, the functional configuration, and the morphological analysis and the retrieval expression creating process in the input analysis unit 12 of the seventh embodiment are the same as those in the first embodiment described above, and the description thereof will be omitted. The description of the document data storage unit 17, the search unit 15, the search result display unit 16 and the like is the same as in the first embodiment described above, and therefore will be omitted here.

The retrieval unit 15 collates the character string in the document of the retrieval result with the input character string. And FIG.
As shown in 2, the search result display section including the input character string is displayed.

In FIG. 42, the result of searching for a sentence containing the character string "calculator" from the document of the search result when "calculator" is input is displayed. Also, "Calculator"
About the character string immediately after, Hiragana, Katakana, Kanji,
According to the other four character types, only a predetermined one sentence may be displayed as the sentence having the same characters in the range of the same character type in the character string immediately after “Computer”. FIG. 42
43 shows the result of performing this process.

As described above, according to the seventh embodiment, it is possible to display what kind of character string the character string that matches the search word is adjacent to. Therefore, it is possible to obtain an effect of preventing omission of the search.

The dictionaries of the first to fourth embodiments described above are not limited to the dictionaries for the retrieval process, but can also be synonymous candidates using the user's glossary, dictionaries such as machine translation systems and word processors. Can be extended to create.

In the fifth, sixth, and seventh embodiments described above, the case where the word following the search word is displayed has been described, but the word following the search word or both of them are searched,
It is possible to display it. In the first to seventh embodiments, examples of Japanese analysis and Japanese document retrieval have been described, but it is possible to use other languages as well.

[0094]

As described in detail above, according to the present invention, when a search request sentence is input, if the user requests it, the Japanese analysis dictionary and the document to be searched are searched for synonyms. Since it is displayed, etc., it is possible to let the user know what kind of search term / search formula should be input, and it is possible to obtain search results with less omission of search with less effort. .

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a functional configuration of the first embodiment.

FIG. 3 is a diagram showing an example of an analysis result of an input analysis unit of the first embodiment.

FIG. 4 is a diagram showing an example of a display of an input unit of the first embodiment.

FIG. 5 is a diagram showing an example of a display of an input unit of the first embodiment.

FIG. 6 is a diagram showing a format of an analysis dictionary according to the first embodiment and an example thereof.

FIG. 7 is a diagram showing an example of a storage format of an analysis dictionary of the first embodiment.

FIG. 8 is a flowchart showing a search process of an analysis dictionary according to the first embodiment.

FIG. 9 is a diagram showing an example of a display of an input unit of the first embodiment.

FIG. 10 is a diagram showing an example of a display of the input unit of the first embodiment.

FIG. 11 is a diagram showing a keyword index table of the first embodiment.

FIG. 12 is a flowchart showing the processing of the search unit of the first embodiment.

FIG. 13 is a diagram showing a format of an analysis dictionary according to a second embodiment of the present invention and an example thereof.

FIG. 14 is a diagram showing an example of an analysis result of the input analysis unit of the second embodiment.

FIG. 15 is a diagram showing an example of an unnecessary expression rule according to the second embodiment.

FIG. 16 is a diagram showing an example of a display of an input unit of the second embodiment.

FIG. 17 is a diagram showing an example of a display on the input unit according to the second embodiment.

FIG. 18 is a flowchart showing the search processing of the analysis dictionary of the second embodiment.

FIG. 19 is a diagram showing an example of a display on the input unit according to the second embodiment.

FIG. 20 is a diagram showing an example of a display of an input unit of the second embodiment.

FIG. 21 is a diagram showing an example of a display of an input unit according to the third embodiment of the present invention.

FIG. 22 is a diagram showing an example of a display on the input unit of the third embodiment.

FIG. 23 is a flowchart showing the search processing of the analysis dictionary of the third embodiment.

FIG. 24 is a diagram showing an example of a display of an input unit of the third embodiment.

FIG. 25 is a diagram showing an example of a display of an input unit of the third embodiment.

FIG. 26 is a diagram showing an example of a display of an input unit according to the fourth embodiment of the present invention.

FIG. 27 is a diagram showing an example of a display on the input unit according to the fourth embodiment.

FIG. 28 is a flowchart showing the search processing of the analysis dictionary of the fourth embodiment.

FIG. 29 is a diagram showing an example of a display on the input unit according to the fourth embodiment.

FIG. 30 is a diagram showing an example of a display on the input unit of the fourth embodiment.

FIG. 31 is a diagram showing an example of a display of an input unit according to the fifth embodiment of the present invention.

FIG. 32 is a diagram showing an example of a display on the input unit of the fifth embodiment.

FIG. 33 is a diagram showing a keyword index table of the fifth embodiment.

FIG. 34 is a diagram showing an example of a display on the input unit of the fifth embodiment.

FIG. 35 is a diagram showing an example of a display on the input unit of the fifth embodiment.

FIG. 36 is a diagram showing an example of a display of an input unit according to the sixth embodiment of the present invention.

FIG. 37 is a diagram showing an example of a display on the input unit of the sixth embodiment.

FIG. 38 is a diagram showing an example of a display on the input unit according to the sixth embodiment.

FIG. 39 is a diagram showing a semantic structure index table according to the sixth embodiment.

FIG. 40 is a diagram showing a semantic structure index table according to the sixth embodiment.

FIG. 41 is a diagram showing an example of a display on the input unit according to the sixth embodiment.

FIG. 42 is a diagram showing an example of a display on the input unit according to the seventh embodiment of the present invention.

FIG. 43 is a diagram showing an example of a display on the display unit of the seventh embodiment.

FIG. 44 is a diagram showing a format of an analysis dictionary and an example thereof.

FIG. 45 is a flowchart showing a search process of an analysis dictionary.
To.

FIG. 46 is a flowchart showing a search process of an analysis dictionary.
To.

[Explanation of symbols]

1 ... Central processing unit, 2 ... Storage unit, 3 ... Display controller
La, 4 ... Display unit, 5 ... Input controller, 6, 11 ... Input unit, 12 ... Input analysis unit, 13 ... Analysis dictionary, 14 ... Control unit, 15 ... Search unit, 16 ... Search result display unit, 17 ... Document data storage unit.

Claims (6)

[Claims] 1. A dictionary that stores at least headwords, a display unit, a document storage unit that stores a plurality of documents, an input unit that inputs a search character string, and an input by this input unit. A dictionary search means for searching the dictionary for a headword composed of a character string including a search character string, a means for displaying the search result searched by the dictionary search means on the display means, and a display means for displaying on the display section. Selecting means for selecting a search character from the search results, searching means for searching the search character selected by the selecting means from the documents stored in the document storing means, and search performed by the searching means A document retrieving apparatus comprising: means for displaying the result on the display means. 2. A dictionary that stores at least headwords, a display unit, a document storage unit that stores a plurality of documents, an input unit that inputs a search character string, and an input by this input unit. A dictionary search unit that searches the dictionary for a headword that matches a part of a search character string, a unit that displays the search results searched by the dictionary search unit on the display unit, and a search displayed on the display unit. Selecting means for selecting a search character from the results; searching means for searching the search character selected by the selecting means from the documents stored in the document storing means; and a search result searched by the searching means. A document retrieval device comprising: means for displaying on the display means. 3. A dictionary for storing word information constituting a compound word, a display means, a document storing means for storing a plurality of documents, an input means for inputting a search compound word, and an input means for inputting by this input means. A dictionary search means for searching the dictionary for a compound word composed of the same constituent words as the constituent word constituting the search compound word; and means for displaying the search result searched by the dictionary search means on the display means. Selecting means for selecting a search character from the search results displayed on the display section; and searching means for searching the document stored in the document storing means for the search character selected by the selecting means; A document retrieving apparatus comprising: means for displaying the retrieval result retrieved by the retrieval means on the display means. 4. A display means, a document storage means for storing a plurality of documents, an input means for inputting a search character string, and a character string adjacent to the search character string input by the input means, in the document. First search means for searching the documents stored in the storage means, means for displaying the search results searched by the search means on the display means, and search characters from the search results displayed on the display section. Selecting means for selecting, the second searching means for searching the search character selected by the selecting means from the documents stored in the document storing means, and the search searched by the second searching means. A document retrieving apparatus comprising: means for displaying the result on the display means. 5. The document searching apparatus according to claim 4, wherein the first searching means searches for a document selected by the user by the selecting means or discarded. 6. The first search means displays only the characters in the same character type range but different from each other immediately before or after the character string input by the user from the input means. The document search device according to claim 4.