JP2008171109A - Information retrieval system, information retrieval method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To display a network diagram showing the relationship between unique expressions (person name or organization name or the like) related to a retrieval key in a short time after the input of the retrieval key. <P>SOLUTION: In the case of registering a document to be retrieved, a relation analyzing part 2 extracts unique expressions included in the document, and generates relation information showing the relationship between the extracted unique expressions, and a relation information registration part 3 registers the relation information generated by the relation analyzing part 2 in a relation index storage device 5. In retrieving the document, when a user inputs a retrieval key, a retrieval part 7 retrieves the relation information adapted to the retrieval key from the relation index storage device 5, and a retrieval result integration part 8 perform the graphic display of the relationship between the unique expressions shown by the relation information retrieved by the retrieval part 7 by using a network diagram. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an information search technique for searching a document, and more particularly to an information search technique that allows a user to easily recognize the relationship between a person, an organization, a product, and the like related to a search key input when searching for a document.

  A conventional general information search system for searching documents includes a document storage unit in which a search target document is registered, and an index storage unit in which a keyword and a document ID of the search target document including the keyword are registered. ing. When the search key is input from the user, the document ID of the search target document including the same keyword as the search key is searched from the index storage unit, and the list is returned to the user. When the document ID is designated by the above, a document with the document ID is retrieved from the document storage unit and returned to the user.

  As described above, when a search key is input, the conventional general information search system simply returns a list of document IDs of documents including the search key to the user. For this reason, the user must actually read the document in order to recognize the relationship between the person, organization, product, and the like related to the search key.

  In order to solve such problems, the following information search system has been proposed (for example, see Patent Document 1).

  The information search system described in Patent Document 1 is a document database in which a plurality of search target documents are registered, and for each search target document, a unique expression that appears in the document and a document ID that are registered in association with each other. And an expression database. Here, the specific expression is a generic name of numerical expressions associated with units such as proper nouns included in categories such as person names, organization names, product names, and place names, and time and ratio.

  Then, the information search system described in Patent Document 1 searches a plurality of documents (conforming documents) that match the search key from the document database when a search key is input by the user, and further, documents for each conforming document. By searching the specific expression database by ID, all the specific expressions included in each relevant document are extracted. Next, the degree of association between each extracted unique expression and the corresponding document group is calculated, and a related unique expression list including a specific expression having a degree of association equal to or greater than a predetermined value is created. After that, the co-occurrence relationship of the specific expressions included in the related specific expression list is obtained, and a graph in which the specific expressions (person name, company name, product name, etc.) are used as nodes and the specific expressions in the co-occurrence relation are connected by arcs. indicate.

Japanese Patent Laid-Open No. 2005-50135

  According to the information search system described in Patent Document 1 described above, the degree of relevance with the relevant document group among the unique expressions that appear in the relevant document group that matches the search key when the user inputs the search key. Since the relationship between the unique expressions (person name, company name, product name, etc.) with high is displayed by the graph, the user does not read the document, but between the person name, company name, product name, etc. related to the search key It becomes possible to recognize the relationship.

  However, in the information search system described in Patent Document 1, each match that matches the search key after the search key is input by the user until the graph representing the co-occurrence relationship between the specific expressions is displayed. Since the process of calculating the relevance between the specific expressions that appear in the document and the matching document group and the process of obtaining the co-occurrence relationship between the specific expressions whose relevance is a predetermined value or more are performed, There was a problem that it took time to recognize the relationship between the company name, product name, and the like.

  In addition, the information search system described in Patent Document 1 displays the co-occurrence relationship between the unique expressions as the relation contents, but it is concrete that the contents of the connection between the co-occurring specific expressions are not read. Has the problem of not knowing.

(Object of invention)
Therefore, a first object of the present invention is to allow a user to recognize the relationship between a person name, a company name, a product name, and the like related to a search key in a short time after inputting the search key.

  A second object of the present invention is to enable the user to recognize the relationship between the search key, such as a person name, company name, and product name, and the contents of the relationship without reading the document.

The first information retrieval system according to the present invention is:
A relationship information storage unit in which relationship information indicating the relationship between specific expressions is registered;
A relation analysis unit that extracts a specific expression included in the document at the time of registration of the document, and generates relation information indicating a relation between the extracted specific expressions;
A registration unit for registering the relationship information generated by the relationship analysis unit in the relationship information storage unit;
A search unit that searches the related information storage unit for related information that matches the search key;
And a search result integration unit that graphically displays the relationship between the unique expressions indicated by the relationship information searched by the search unit.

A second information search system according to the present invention is the first information search system,
The relation information includes a related specific expression and a relation content between the specific expressions.

A third information retrieval system according to the present invention is the second information retrieval system,
The relation analysis unit performs a unified process of notation of a specific expression included in the generated relation information.

A fourth information search system according to the present invention is the second information search system,
The search result integration unit represents a specific expression included as a subject or object in each relationship information searched by the search unit by a node, and a relationship between the specific expressions is expressed by an arc connecting the nodes and a relationship content. A network diagram represented by a character string to be displayed is displayed.

A fifth information search system according to the present invention is the fourth information search system,
A network diagram in which the search result integration unit represents a relationship between specific expressions by an arc connecting nodes, a character string indicating the relationship content, and a character string indicating geographic information or time information for modifying the relationship content. Is displayed.

A sixth information search system according to the present invention is the first information search system,
The search unit is configured to search for relation information using a search key related to a specific expression included in the searched relation information for the number of hierarchies designated by the user.

A first information search method according to the present invention is as follows.
A method of searching for information using a computer,
An extraction step in which the computer extracts a specific expression included in the document when registering the document;
A relation information generation step in which the computer generates relation information indicating a relation between the unique expressions extracted in the extraction step;
A registration step in which the computer registers the relationship information generated in the relationship information generation step in a relationship information storage unit;
A search step in which the computer searches the relationship information storage unit for relationship information that matches a search key;
The computer further includes a display step for graphically displaying the relationship between the unique expressions indicated by the relationship information searched in the search step.

A second information retrieval method according to the present invention is the first information retrieval method,
The relation information includes a related specific expression and a relation content between the specific expressions.

A third information retrieval method according to the present invention is the second information retrieval method,
In the relationship information generation step, the notation of the unique expression included in the generated relationship information is performed.

A fourth information search method according to the present invention is the second information search method,
In the display step, a specific expression included as a subject or object in each relationship information searched in the search step is represented by a node, and a relationship between the specific expressions is represented by an arc connecting the nodes and a character indicating the content of the relationship. A network diagram represented by columns is displayed.

A fifth information retrieval method according to the present invention is the fourth information retrieval method,
In the display step, a network diagram is displayed in which the relationship between the unique expressions is represented by an arc connecting the nodes, a character string indicating the relationship content, and a character string indicating the geographic information or time information for modifying the relationship content. It is characterized by doing.

A sixth information retrieval method according to the present invention is the first information retrieval method,
In the search step, the relationship information search using a search key related to the specific expression included in the searched relationship information is performed for the number of hierarchies designated by the user.

The first program according to the present invention is:
A computer including a relationship information storage unit in which relationship information indicating the relationship between specific expressions is registered,
A relation analysis unit that extracts a specific expression included in the document and generates relation information indicating a relation between the extracted specific expressions when the document is registered;
A registration unit for registering the relationship information generated by the relationship analysis unit in the relationship information storage unit;
A search unit for searching related information matching the search key from the related information storage unit;
It is made to function as a search result integration unit that graphically displays the relationship between the unique expressions indicated by the relationship information searched by the search unit.

The second program according to the present invention is the first program,
The relation information includes a related specific expression and a relation content between the specific expressions.

A third program according to the present invention is the second program,
The relation analysis unit performs a unified process of notation of a specific expression included in the generated relation information.

A fourth program according to the present invention is the second program,
The search result integration unit represents a specific expression included as a subject or object in each relationship information searched by the search unit by a node, and a relationship between the specific expressions is expressed by an arc connecting the nodes and a relationship content. A network diagram represented by a character string to be displayed is displayed.

A fifth program according to the present invention is the fourth program,
A network diagram in which the search result integration unit represents a relationship between specific expressions by an arc connecting nodes, a character string indicating the relationship content, and a character string indicating geographic information or time information for modifying the relationship content. Is displayed.

A sixth program according to the present invention is the first program,
The search unit is configured to search for relation information using a search key related to a specific expression included in the searched relation information for the number of hierarchies designated by the user.

[Action]
When registering a document to be searched, the relationship analysis unit extracts the specific expressions included in the document, generates relationship information indicating the relationship between the extracted specific expressions, and the registration unit displays the relationship analysis unit. The relationship information generated in step 1 is registered in the relationship information storage unit. Note that the relationship information includes, for example, related unique expressions and the related contents.

  When a user inputs a search key when searching for a document, the search unit searches the related information storage unit for related information that matches the search key, and the search result integrating unit is indicated by the related information searched by the search unit. The relationship between the specific expressions is displayed graphically using, for example, a network diagram.

  According to the present invention, it is possible to recognize a relationship such as a person, an organization, and a product (represented by a unique expression) related to a search key in a short time after the user inputs the search key. The reason is that, when registering a document, relationship information indicating the relationship between each unique expression included in the document is generated and registered in the relationship information storage unit. That is, according to the present invention, when a search key is input, the relationship information storage unit is searched for the relationship information that matches the search key, and the relationship between the specific expressions indicated by the searched relationship information is displayed graphically. Since the search key is entered, processing to calculate the relevance between the specific expressions that appear in each relevant document that matches the search key and the relevant document group, and between the specific expressions whose relevance is greater than or equal to the specified value. Compared to the information retrieval system described in Patent Document 1 that had to perform the process of obtaining the co-occurrence relationship, the relationship between the unique expressions can be displayed in a graphic in a short time, and as a result, the user can Can recognize the relationship between a person, an organization, a product, and the like related to the search key in a short time.

  Furthermore, according to the present invention, it becomes possible for the user to recognize the relationship between the search key, such as a person name, company name, and product name, and the content of the relationship without reading the document. The reason is that the relationship between the unique expressions is graphically displayed based on the relationship information including the related specific expressions and the relationship contents between the specific expressions.

  Next, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

[Description of Configuration of Embodiment]
Referring to FIG. 1, an information retrieval system 100 according to an embodiment of the present invention includes a registered document input unit 1, a relationship analysis unit 2, a relationship information registration unit 3, a document registration unit 4, and a relationship index storage device. 5, a document storage device 6, a search unit 7, a search result integration unit 8, an input unit 9 such as a keyboard, and a display unit 10 such as an LCD.

  The registered document input unit 1 has a function of inputting a document described in a natural language to be searched.

  The document registration unit 4 gives the document input from the registered document input unit 1 a unique document ID and registers it in the document storage device 6, and notifies the related information registration unit 3 of the document ID assigned to the document. It has the function to do.

  The relationship analysis unit 2 analyzes the document input from the registered document input unit 1, and shows relationship information indicating the relationship between specific expressions such as a person name, organization name, product name, place name, and time included in the document. And the generated relationship information is passed to the relationship information registration unit 3.

  Referring to FIG. 2, a detailed configuration example of the relationship analysis unit 2 illustrated in FIG. 1 is illustrated. A natural language analysis unit 21, a specific expression extraction unit 22, a relationship extraction unit 23, a notation unification unit 24, and the like. It has.

  The natural language analysis unit 21 analyzes each sentence described in the natural language in the document input from the registered document input unit 1, determines the dependency structure of the sentence, and resolves the pronoun / noun abbreviation. The natural language analysis unit 21 performs basic analysis of natural language processing, such as morphological analysis, syntax analysis, semantic analysis, and context analysis. Since each analysis method is not directly related to the present invention, its detailed configuration is omitted.

  The specific expression extraction unit 22 analyzes each sentence in the document input from the registered document input unit 1 and extracts a specific expression. The specific expression extraction unit 22 extracts a specific expression using an arbitrary extraction rule such as an extraction rule created manually or an extraction rule created by machine learning (maximum entropy method, hidden Markov model, support vector machine, etc.). To do. Since the specific expression extraction method is not directly related to the present invention, its detailed configuration is omitted.

  The relationship extraction unit 23 is provided by the natural language analysis unit 21, the dependency structure information of the sentence, the information of demonstrative pronouns and abbreviations that are replaced / inserted into specific words, and the specific expression extraction unit 22. Using the unique expression information in the sentence, relation information indicating the relationship between the unique expressions is generated.

  The notation unifying unit 24 performs unification processing of the unique expression included in the relationship information generated by the relationship extracting unit 23. More specifically, unification processing of specific expressions representing time and unification processing of synonyms are performed. Here, as a process for unifying proper expressions representing time, for example, there is a process for unifying year, month and day in the Western calendar. The synonym unifying process includes, for example, a process of unifying an abbreviation such as “JOC” into an official name such as “Japan Olympic Committee”.

  FIG. 3 illustrates the relationship information indicating the relationship between the specific expressions. FIG. 3 illustrates three pieces of relationship information. In the example of FIG. 3, assuming that the specific expressions included in the categories “place name” and “time expression” are extracted, the specific expressions of other categories, Named_Entity1 (subject specific expression) and Named_Entity2 ( It indicates that the specific expression (object) is related by an action indicating an action relationship, and further indicates that the action operation date is Datetime and the action location is a location. The content of the relationship to be extracted is not limited to the format shown in FIG. 3 as long as it represents the relationship between the specific expressions.

  Referring back to FIG. 1 again, the relationship information registration unit 3 registers the relationship information passed from the relationship analysis unit 2 and the document ID passed from the document registration unit 4 in the relationship index storage device 5.

  In the relation index storage device 5, a pair of relation information and a document ID of a document used when generating the relation information is registered. FIG. 4 shows an example of the contents of the relation index storage device 5. The first line of the example in the figure includes related information indicating that Vincent Clark belonged to NNA Org. On 04/01/2004, and this related information is included in the document with the document ID “ID1”. It is generated based on this.

  A plurality of documents to be searched are registered in the document storage unit 6. Each document is given a unique document ID.

  The search unit 7 searches the relationship index storage device 5 for a pair of relationship information and document ID that matches the search key input from the input unit 9 and searches for the relationship information included in each searched pair. It has a function to pass to the result integration unit 8 and a function to display a list of document IDs included in each searched pair on the display unit 10. Further, the search unit 7 has a function of searching the document storage device 6 for the document with the document ID input from the input unit 9 and displaying the document on the display unit 10.

  The search result integration unit 8 has a function of analyzing each relationship information (search result) given from the search unit 7 and displaying a network diagram representing the relationship between the unique expressions indicated by each relationship information on the display unit 10. Have.

  Note that the information search system 100 can be realized by a computer, and when realized by a computer, for example, is as follows. A disk, a semiconductor memory, and other recording media on which a program for causing a computer to function as the information retrieval system 100 is recorded are prepared, and the computer is caused to read the program. The computer controls the operation of the computer according to the read program, thereby realizing the relationship analysis unit 2, the relationship information registration unit 3, the document registration unit 4, the search unit 7, and the search result integration unit 8 on the computer. .

[Description of Operation of Embodiment]
Next, the operation of the present embodiment will be described in detail.

[Operation when registering a document]
First, an operation when registering a document to be searched will be described. The administrator of the information search system 100 inputs a document from the registered document input unit 1 when registering a document to be searched.

  Thereby, as shown in the flowchart of FIG. 5, the document registration unit 4 assigns a unique document ID to the input document and registers it in the document storage device 6 (steps S51 and S52), and then assigns it to the document. The related document ID is notified to the related information registration unit 3 (step S53).

  When the document is input from the registered document input unit 1, the relationship analysis unit 2 pays attention to one sentence in the input document as shown in the flowchart of FIG. 6 (step S61). After that, it generates relationship information indicating the relationship that exists between the specific names such as person name, organization name, product name, place name, date and time included in the sentence of interest, and the specific name in the related information And the related information after the unified processing is passed to the related information registration unit 3 (steps S63 and S64). FIG. 3 shows an example of the relationship information generated. In FIG. 3, one relation information is shown in one line, and the first line is "ABC Corp. (organization name)" to "Mr. Vincent Clark (person name)" It shows that there is a relationship of “greeting” at the time of (date).

  The relationship analysis unit 2 repeatedly performs the above-described processing until there is no unfocused sentence. When there is no unfocused sentence (YES in step S62), an end notification is sent to the relationship information registration unit 3 (step S65). ).

  Further, when the document is input from the registered document input unit 1, the relationship information registration unit 3 waits for input of a document ID, end notification, or relationship information as shown in the flowchart of FIG. 7 (steps S71 to S71). S73).

  If a document ID is input from the document registration unit 4 (YES in step S71), the document ID is stored in a document ID holding unit (not shown) (step S76), and then the process of step S71 is performed. Return.

  When the relationship information is passed from the relationship analysis unit 2 (YES in step S73), it is checked whether or not the document ID is held in the document ID holding unit (step S74).

  If the document ID is not held (NO in step S74), the relation information input from the relation analysis unit 2 is stored in a relation information holding unit (not shown) (step S77), and then step S71. Return to the process.

  On the other hand, when the document ID is held (YES in step S74), the relation index storage device displays the relation information input from the relation analysis unit 2 and the document ID held in the document ID holding unit. 5 (step S75), and the process of step S71 is performed again. In step S75, if related information is held in the related information holding unit, the related information is also registered in the related index storage device 5 together with the document ID.

  If an end notification is input from the relationship analysis unit 2 (YES in step S72), the relationship information registration unit 3 deletes the document ID held in the document ID holding unit (step S79), and then performs processing. Exit.

  Through the above processing, the relationship information related to the document registered this time is registered in the relationship index storage device 5, and the knowledge indicating the relationship between the specific expressions is updated. Now, for example, if the contents of the relationship index storage device 5 are as shown in FIG. 4, the relationship between the unique expressions represented by the relationship information registered in the relationship index storage device 5 is as shown in FIG. As shown in A). In FIG. 8A, the character string on the arc indicating the relationship between the nodes is not shown. Thereafter, when the three pieces of relational information shown in FIG. 3 are registered in the relational index storage device 5 at the time of registration of a new document, between the unique expressions represented by the pieces of relational information registered in the relational index storage unit 5 The relationship is as shown in FIG.

[Operation when searching documents]
Next, the operation at the time of document search will be described.

  When searching for a document, the user inputs a search key from the input unit 9. The search key format can be arbitrary, but in this embodiment, the search key of the format shown in FIG. 9 is used. Referring to FIG. 9, in the search key of this embodiment, the phrase of the specific expression to be searched, the category to which the specific expression to be searched belongs, the contents of the relational operation to be extracted, and the extraction target The category of the specific expression and the time range to be extracted can be specified. Each item can be specified with a plurality of conditions under the OR condition. For example, during the period of “2004/11/12 to 2006/11/12”, “person (person name)” and “organization (name)” related to “person (person name)” “Mr. Vincent Clark” If it is desired to obtain this information, conditions as shown in FIG. 10 are input.

  When the search key is input from the input unit 9, the search unit 7 searches the relationship index storage device 5 for all relation information and document ID pairs that match the search key, as shown in the flowchart of FIG. 11. (Step S111). Thereafter, the search unit 7 displays the searched document ID list in the list display area of the display unit 10, and further passes the searched list of related information to the search result integration unit 8 (steps S112 and S113).

  When a list of relation information is passed from the search section 7, the search result integration section 8 analyzes the relation information included in the list as shown in the flowchart of FIG. A network diagram indicating the relationship is displayed in the network diagram display area of the display unit 10 (steps S121 and S122). In this embodiment, the specific expression that is the subject or object is a node, the operation relationship (relationship content) between the specific representations is an arc (including a character string indicating the content of the operation relationship), and the operation relationship is further modified. A network diagram in which geographical information and time information are arranged on the corresponding arc is displayed.

  Now, for example, if the list of relation information passed from the search unit 7 is as shown in FIG. 13, the search result integration unit 8 displays the network diagram shown in FIG. 14 on the display unit 10. By referring to such a network diagram, the user can easily confirm the relationship between persons, organizations, and the like related to the search key.

  Further, when referring to a document included in the list of document IDs, the user inputs the document ID of the document to be referred to from the input unit 9. When the document ID is input, the search unit 7 searches the document storage device 6 for the document with the document ID and displays it on the display unit 10 as shown in the flowchart of FIG. 15 (step S151, S152).

[Effect of the embodiment]
According to the present embodiment, it is possible to recognize a relationship such as a person, an organization, or a product (represented by a specific expression) related to the search key in a short time after the user inputs the search key. Become. The reason is that, when the document is registered, the relationship analysis unit 2 generates relationship information indicating the relationship between the unique expressions included in the document, and the relationship information registration unit 3 is generated by the relationship analysis unit 2. This is because the relationship information is registered in the relationship index storage device 5. That is, in this embodiment, when a search key is input, the search unit 7 performs a process of searching the related index storage device 5 for related information that matches the search key, and the search result integrating unit 8 is searched. Since it is only necessary to perform a graphic display of the relationship between the specific expressions indicated by the relationship information, after the search key is entered, the relationship between the specific expressions that appear in each relevant document that matches the search key and the relevant document group Compared to the information retrieval system described in Patent Document 1 in which processing for calculating the degree and processing for obtaining a co-occurrence relationship between specific expressions having a degree of association equal to or greater than a predetermined value must be performed. The relationship between expressions can be displayed graphically. As a result, the user can recognize the relationship between a person, an organization, a product, and the like related to the search key in a short time.

  Furthermore, according to the present embodiment, it becomes possible for the user to recognize the relationship such as a person name, company name, and product name related to the search key and the content of the relationship without reading the document. The reason is that the relationship between the unique expressions is graphically displayed based on the relationship information including the related specific expressions and the relationship contents between the specific expressions.

  Further, in the present embodiment, since the relationship analysis unit 2 performs the unification processing of the unique expressions in the relationship information, it is possible to relate the unique expressions that are different in the document.

[Other Embodiments]
In the above-described embodiment, the search unit 7 is a search unit that searches for a pair of relation information and document ID matching the search key from the relation index storage device 5 and has one search hierarchy. However, instead, a search unit that performs a search for the number of search hierarchies designated by the user can also be used. That is, if the search is performed on the first layer using the search key input from the input unit 9 by the user, and if the search is not performed for the number of layers specified by the user, the relationship retrieved in the first layer is further included. Use a search unit that keeps searching until the number of search hierarchies specified by the user is searched, such as a search of the second hierarchy using a search key related to the specific expression included in the information You may do it.

[Effects of other embodiments]
In this embodiment, since a search unit that performs a search for the number of search hierarchies specified by the user is used, not only the relationship between the specific expressions related to the search key input by the user, but also the relationship of the above specific expressions It is also possible to graphically display the relationship between the unique expressions.

  The present invention can be applied to an information retrieval system such as a document retrieval system that retrieves a document.

It is a block diagram which shows the structural example of embodiment of the information search system concerning this invention. 3 is a block diagram illustrating a configuration example of a relationship analysis unit 2. FIG. It is a figure which shows an example of relationship information. It is a figure which shows the example of the content of the relationship index memory | storage device 5. FIG. 6 is a flowchart illustrating a processing example of a document registration unit 4. 5 is a flowchart illustrating a processing example of a relationship analysis unit 2. 5 is a flowchart illustrating a processing example of a relationship information registration unit 3. It is a figure which shows that the knowledge showing the relationship between specific expressions is accumulate | stored by registering relationship information in the relationship index memory | storage device 5. FIG. It is a figure which shows an example of the format of a search key. It is a figure which shows the specific example of a search key. It is a flowchart which shows the process example of the search part 7 when a search key is input. 10 is a flowchart illustrating a processing example of a search result integration unit 8. It is a figure which shows an example of the relationship information searched in the search part. It is a network diagram showing the relationship between the specific expressions shown by the relationship information of FIG. It is a figure which shows the example of a process of the search part 7 when document ID is input.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Registered document input part 2 ... Relation analysis part 21 ... Natural language analysis part 22 ... Specific expression extraction part 23 ... Relation extraction part 24 ... Notation unification part 3 ... Relation information registration part 4 ... Document registration part 5 ... Relation index memory | storage device 6 ... Document storage device 7 ... Search unit 8 ... Search result integration unit 9 ... Input unit 10 ... Display unit 100 ... Information search system

Claims (18)

A relationship information storage unit in which relationship information indicating the relationship between specific expressions is registered;
A relation analysis unit that extracts a specific expression included in the document at the time of registration of the document, and generates relation information indicating a relation between the extracted specific expressions;
A registration unit for registering the relationship information generated by the relationship analysis unit in the relationship information storage unit;
A search unit that searches the related information storage unit for related information that matches the search key;
An information search system comprising: a search result integration unit that graphically displays a relationship between specific expressions indicated by the relationship information searched by the search unit. The information search system according to claim 1,
The information search system, wherein the relation information includes a related specific expression and a relation content between the specific expressions. The information search system according to claim 2,
The information search system, wherein the relation analysis unit performs a unified process of notation of a specific expression included in the generated relation information. The information search system according to claim 2,
The search result integration unit represents a specific expression included as a subject or object in each relationship information searched by the search unit by a node, and a relationship between the specific expressions is expressed by an arc connecting the nodes and a relationship content. An information search system for displaying a network diagram represented by a character string to be displayed. The information search system according to claim 4,
A network diagram in which the search result integration unit represents a relationship between specific expressions by an arc connecting nodes, a character string indicating the relationship content, and a character string indicating geographic information or time information for modifying the relationship content. An information retrieval system characterized by displaying The information search system according to claim 1,
The information search system, wherein the search unit searches for related information using a search key related to a specific expression included in the searched related information for the number of hierarchies designated by the user. A method of searching for information using a computer,
An extraction step in which the computer extracts a specific expression included in the document when registering the document;
A relation information generation step in which the computer generates relation information indicating a relation between the unique expressions extracted in the extraction step;
A registration step in which the computer registers the relationship information generated in the relationship information generation step in a relationship information storage unit;
A search step in which the computer searches the relationship information storage unit for relationship information that matches a search key;
An information search method comprising: a display step in which the computer graphically displays a relationship between specific expressions indicated by the relationship information searched in the search step. The information search method according to claim 7,
The information search method, wherein the relation information includes a related specific expression and a relation content between the specific expressions. The information search method according to claim 8, wherein
In the relation information generation step, an unification process of notation of specific expressions included in the generated relation information is performed. The information search method according to claim 8, wherein
In the display step, a specific expression included as a subject or object in each relationship information searched in the search step is represented by a node, and a relationship between the specific expressions is represented by an arc connecting the nodes and a character indicating the content of the relationship. An information retrieval method, characterized by displaying a network diagram represented by columns. The information search method according to claim 10,
In the display step, a network diagram is displayed in which the relationship between the unique expressions is represented by an arc connecting the nodes, a character string indicating the relationship content, and a character string indicating the geographic information or time information for modifying the relationship content. An information retrieval method characterized by: The information search method according to claim 7,
In the retrieval step, the retrieval of the relation information using the retrieval key related to the specific expression included in the retrieved relation information is performed for the number of hierarchies designated by the user. A computer including a relationship information storage unit in which relationship information indicating the relationship between specific expressions is registered,
A relation analysis unit that extracts a specific expression included in the document and generates relation information indicating a relation between the extracted specific expressions when the document is registered;
A registration unit for registering the relationship information generated by the relationship analysis unit in the relationship information storage unit;
A search unit for searching related information matching the search key from the related information storage unit;
A program for functioning as a search result integration unit that graphically displays the relationship between specific expressions indicated by the relationship information searched by the search unit. The program according to claim 13, wherein
The program characterized in that the relation information includes a related specific expression and a relation content between the specific expressions. The program according to claim 14, wherein
A program characterized in that the relation analysis unit performs a unified process of notation of a specific expression included in the generated relation information. The program according to claim 14, wherein
The search result integration unit represents a specific expression included as a subject or object in each relationship information searched by the search unit by a node, and a relationship between the specific expressions is expressed by an arc connecting the nodes and a relationship content. A program that displays a network diagram represented by a character string to be displayed. The program according to claim 16, wherein
A network diagram in which the search result integration unit represents a relationship between specific expressions by an arc connecting nodes, a character string indicating the relationship content, and a character string indicating geographic information or time information for modifying the relationship content. A program characterized by displaying. The program according to claim 13, wherein
A program characterized in that the search unit searches for relation information using a search key related to a specific expression included in the searched relation information for the number of hierarchies designated by the user.

Images