JPH07311780A - Information retrieval device and information retrieval method - Google Patents

Abstract

PURPOSE:To provide an information retrieval device capable of taking out a document relating to the document defined as an object in a form capable of discriminating the importance degree. CONSTITUTION:This information retrieval device is provided with a disk device 104 for storing a citing relation data base where mutual citing relation among the documents is described and a central processing unit 102 for performing the processings of judgement and arithmetic operations, etc., based on a control program stored in a control memory 101. The central processing unit 102 retrieves the citing relation data base based on a retrieval program, performs the processing for extracting the documents relating to the document defined as the object and performs the processing for ranking the extracted documents in the order of the importance degrees based on a ranking program. Evaluation for the importance degree is performed based on the number of times of citing the document.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieving apparatus and an information retrieving method for retrieving documents related to a target document based on mutual citation relations between documents.

[0002]

2. Description of the Related Art Generally, when searching for a document related to a target document, a procedure for sequentially searching for references cited in the target document is adopted, and the reference documents are listed in enumeration. References that are present are retrieved, but it is unclear whether any of the references is an important reference with a content that provides the basis for the content of the document of interest.

[0003]

However, it is clarified whether any of the references in the target document is an important reference having a content that serves as a basis for the content of the target document. As such, it is difficult to determine which references are important and which are unimportant. Further, in practice, it is not possible to search for a document that is cited as a target document but is not explicitly specified as a reference document, or a document that has been further developed by citing the target document.

An object of the present invention is to provide an information retrieval apparatus and an information retrieval method which can retrieve a document related to a target document so that its importance can be determined.

[0005]

The invention according to claim 1 is
In an information retrieval device that searches for documents related to a target document based on mutual citation relations between documents, a storage unit that holds a citation relation database in which citation relations between documents are described, and the citation relation database are searched. Then
It is characterized by comprising a searching means for extracting documents related to the target document and a ranking means for ranking the extracted documents in order of their importance.

The invention according to claim 2 is the information search apparatus according to claim 1, wherein the importance of the document is determined based on the number of times the document is cited by another document. .

According to a third aspect of the present invention, in the information retrieval apparatus according to the first aspect, the extracted documents are output to the outside in the order of importance.

According to a fourth aspect of the invention, in an information retrieval method for searching for documents related to a target document based on mutual citation relations between documents, a citation relation database in which mutual citation relations between documents are described is retained. And a step of searching the citation relational database to extract documents related to the target document, and a step of ranking the extracted documents in order of their importance.

According to a fifth aspect of the present invention, in the information search method according to the fourth aspect, the importance of the document is determined based on the number of times the document is cited by other documents. .

[0010]

In the information retrieving apparatus according to the first aspect, a citation relation database in which mutual citation relations between documents are described is stored in the storage means, and the citation relation database is searched by the retrieving means to select the target document. Related documents are extracted, and the documents extracted by the ranking means are ranked in order of their importance.

In the information retrieval device according to the second aspect, the importance of the document is determined based on the number of times the document is cited by other documents.

In the information retrieval device according to the third aspect, the extracted documents are output to the outside in the order of importance.

In the information retrieval method according to the fourth aspect, a citation relation database in which mutual citation relations between documents are described is held, the citation relation database is searched, and documents relevant to a target document are extracted. The extracted documents are ranked in order of their importance.

In the information retrieval method according to the fifth aspect, the importance of a document is determined based on the number of times the document is cited by other documents.

[0015]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the information retrieval apparatus of the present invention, and FIG. 2 is a diagram schematically showing the contents of the citation relation database used in the information retrieval apparatus of FIG. .

As shown in FIG. 1, the information retrieving apparatus according to this embodiment includes a disk unit 104 for storing a citation relation database in which citation relations between documents are described,
A central processing unit 102 that performs processing such as determination and calculation based on a control program stored in the control memory 101.

The citation relation database has a network structure as shown in FIG. Each node represents the substance of a document, and the arrows in the figure represent citation relationships. For example, it is shown that the target document A is cited as the reference documents, the document 1, the document 2, the document 3, and the document 5, and is cited as the document 15, the document 16, and the document 19. In this embodiment, the node representing the substance of the document is used in the citation relation database, but a node representing the identifier of the document may be used instead. Disk device 1
Reference numeral 04 constitutes a storage means for storing a citation relation database.

The control program stored in the control memory 101 includes a search program for searching a citation relation database stored in the disk device 104 and extracting a document related to a target document, and the extraction program. And a ranking program for ranking the cited documents in order of their importance. The importance of a document in a ranked program is evaluated based on the number of times the document is cited, ie, the number of parent nodes.

The central processing unit 102 searches the citation relation database based on the search program, performs a process for extracting documents related to the document of interest, and the extracted documents are important based on the ranking program. Perform processing to rank in order. Central processing unit 1
02 constitutes a search means and a ranking means.

The ranked documents are temporarily stored in the memory 103 together with the information indicating the degree of importance thereof.
The documents stored in 03 are arranged in order of importance. The document read from the memory 103 is output means 10
6, the output means 106 displays the read document on the screen and records it on the recording paper.

The central processing unit 102, the control memory 101, the memory 103, the disk device 104 and the output means 106 are mutually connected via a bus 105.

Next, the search operation of the information search apparatus according to this embodiment will be described with reference to the drawings. FIG. 3 shows FIG.
3 is a flowchart showing a search operation of the information search device of FIG.

When searching for a document related to the target document A, the start of the search operation is instructed. When the search operation instruction is input, as shown in FIG. 3, the list L in the area reserved in the memory 103 is emptied (step S3).
1) The user's search request C is input (step S3).
2). In this embodiment, two request examples will be described as the search request C. One of the search requests is a search request C1 of a past document related to the document A, and the other one is a search request C2 of a document related to the document A and published after the document A.

When the search request C is the search request C1, a set Ω of past documents related to the document A is created (step S33). As shown in FIG. 2, this set Ω is composed of a set of subordinate nodes in the node of document A, and is represented by the following equation (1).

[0026]

[Equation 1] Ω = {1, 2, 3, 4, 5, 6, 8.10, 11, 13} (1) Next, an arbitrary element E of the set Ω, that is, the document E is extracted (step S35). The number n of documents that cite this document E is counted (step S36). This document E
The number n of documents quoting is the number of parent nodes in the node of document E, as shown in FIG.

After counting the number n of documents that cite the document E, the data (E, n) including the document E and the number n of its parent nodes is added to the list L (step S37).
Sorting is performed in descending order of n (step S38).

After sorting, the document E is deleted from the set Ω (step S39). Next, it is determined whether or not the set Ω is empty (step S40). If the set Ω is not empty, the process from step S35 is repeated again.
When the set Ω becomes empty, the search result data is output by the output means 106.
Is output to, and the search operation ends.

Through the series of processes described above, the related documents are described in the list L in descending order of the number of times they are cited, the description content of the list L is given to the output means 106 as the search result, and the search result is displayed. Will be recorded. Search results are
It becomes a data string of "Reference 8, Reference 13, Reference 5, Reference 6, Reference 10, ..." Arranged in order from the largest number of parent nodes.

From these search results, the documents 8 and 13 are regarded by the user as documents (eg, the ones that have given the idea that forms the basis of document A) important for understanding document A.

Next, when a search request C2 is issued, the document A
A set Ω of documents related to the document A and published after the document A is created (step S34). This set Ω is a set of nodes having the document A as a lower node, and the set Ω is represented by the following equation (2).

[0032]

[Equation 2] Ω = {15, 16, ...} (2) Next, similarly to the search request C1, the processes from step S35 to step S39 are sequentially performed. The output search result is a data string of “reference 15, reference 16, ...”, arranged in descending order of the number of parent nodes.

From these search results, it is presumed that Documents 15, 16 and the like are documents whose contents are further developed by referring to Document A (for example, documents to which the idea which forms the basis of Document A is given).

From the above, among the documents cited or not cited in the document A, and the documents before or after the issuance time of the document A, the documents related to the document A are considered to be important. Can be extracted in a distinguishable manner.

In this embodiment, the set of related documents of the target document A is created for all the documents in the citation relation database. However, when the citation relation database is large, the set of related documents is set. Can be large. Therefore, the number of related documents to be extracted can be limited and the scale of the set Ω can be reduced by providing the search conditions for the documents related to the target document A. For example, it is possible to reduce the set scale of the extracted documents by designating “related documents within the past 10 years with respect to the document A” or the like.

Although the importance is evaluated based on the number of times it is quoted (the number of parent nodes), other evaluation methods can be used instead. For example, other evaluation methods include a heuristic method and a method using quantitative data.

Further, as the search results of the related documents, the search results in which the related documents are simply listed in the order of importance are presented to the user. It is also possible to set a method of extracting information and set to present a search result in which the information is described to the user.
For example, both the document 8 and the document 13 shown in FIG. 2 can be regarded as documents that are frequently cited, but the document 8 is often cited in the related document of the document A, whereas , Document 13 is often cited by documents other than the related document of Document A. Therefore, Document 8 is a document related to the same field as Document A (for example, a document related to information retrieval), and Document 13 is a general-purpose elemental technology with high field independence (for example, neural network, fuzzy, etc.).
It can be presumed to be a document relating to, and such information can be presented to the user.

In addition, even in a document that cites the document A, when a large number of other documents related to the document A are cited and the number of the cited documents is large, a survey of the same field as the document A is conducted. This information can be presented to the user.

[0039]

As described above, according to the information retrieving apparatus of the first aspect, the citation relation database in which the citation relations between documents are described is retrieved, and the documents related to the subject document are searched. Is extracted and the extracted documents are ranked in the order of their importance, so that it is possible to extract the documents related to the target document so that their importance can be determined.

According to the information retrieving apparatus of the second aspect, the importance of the document is determined based on the number of times the document is cited by other documents, so that the importance of the document can be easily determined. .

According to the information retrieval apparatus of the third aspect, the extracted documents are output to the outside in the order of their importance.
The importance of the document can be easily grasped.

According to the information retrieval method of the fourth aspect, the citation relation database in which the citation relations between documents are described is held, the citation relation database is searched, and the documents related to the target document are extracted. Then, since the extracted documents are ranked in the order of their importance, the documents related to the target document can be extracted so that their importance can be determined.

According to the information retrieval method of the fifth aspect, since the importance of a document is determined based on the number of times the document is cited by another document, the importance of the document can be easily determined. .

[Brief description of drawings]

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

FIG. 2 is a diagram schematically showing the contents of a citation relation database used in the information search device of FIG. 5 is a flowchart showing a search operation of.

FIG. 3 is a flowchart showing a search operation of the information search device of FIG.

[Explanation of symbols]

 101 Control Memory 102 Central Processing Unit 103 Memory 104 Disk Unit 106 Output Means

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Takaya Ueda 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Yuji Ikeda 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Incorporated (72) Inventor Minoru Fujita 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc.

Claims (5)

[Claims] 1. An information search apparatus for searching for documents related to a target document based on mutual citation relations between documents, and storage means for holding a citation relation database in which mutual citation relations between documents are described, An information retrieval device comprising: a retrieval means for retrieving a citation relation database and extracting documents related to the target document; and a ranking means for ranking the extracted documents in order of their importance. . 2. The information retrieval apparatus according to claim 1, wherein the degree of importance of the document is determined based on the number of times the document is cited by another document. 3. The information retrieving apparatus according to claim 1, wherein the extracted documents are output to the outside in the order of their importance. 4. A method for retrieving documents related to a target document based on mutual citation relations between documents, a step of holding a citation relation database in which citation relations between documents are described, and the citation. An information retrieval method comprising: a step of searching a relational database to extract documents related to the target document; and a step of ranking the extracted documents in order of their importance. 5. The information retrieval method according to claim 4, wherein the importance of the document is determined based on the number of times the document is cited by another document.