# US 2010/0138426 A1 · Index generating system, information retrieval system, and index generating method

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2010/0138426 A1
- Document ID: US20100138426A1
- Status: Published application
- Kind: A1
- Application number: 12/626,779
- Filed: 2009-11-27
- Published: 2010-06-03
- Earliest claimed: 2008-12-01
- Assignees: Ntt Docomo, Inc.
- Inventors: Takehiro NAKAYAMA; Daisuke Torii
- Web page: https://patentgrep.com/patent/US20100138426A1

## CPC

- G06F 16/313: {Selection or weighting of terms for indexing}
- G06F 16/334: Query execution (filtering based on additional data G06F16/335)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

In an index generating device, a similarity calculating unit calculates access similarities indicating similarities of access histories between documents, based on history information indicating the access histories to the respective documents by users, and a similar document specifying unit specifies a similar document similar to a given document as a retrieval target, based on the access similarities or the like. A retrieval index generating unit generates a retrieval index for the given document as the retrieval target from words appearing in a document set consisting of the similar document and the given document as the retrieval target.

## Claims

### Claim 1

An index generating system for generating a retrieval index for document retrieval, comprising:
an access history storage unit configured to store history information indicating access histories to respective documents by users;
a similarity calculating unit configured to calculate access similarities indicating similarities of the access histories between the documents, based on the history information stored in the access history storage unit;
a similar document specifying unit configured to specify a similar document similar to a given document as a retrieval target, based on the access similarities; and
a retrieval index generating unit configured to generate a retrieval index for the given document as the retrieval target from words appearing in a document set consisting of the given document as the retrieval target and the similar document.

### Claim 2 (depends on 1)

The index generating system according to claim 1,
wherein the access history storage unit is configured to store as the history information, frequency information wherein access frequencies to the respective documents by the respective users are correlated with identification information of the users,
wherein the similarity calculating unit is configured to calculate the access similarities, based on the frequency information stored in the access history storage unit, and
wherein the similar document specifying unit is configured to compare the access similarities with a similarity determination threshold to specify the similar document.

### Claim 3 (depends on 2)

The index generating system according to claim 2, comprising a link distance calculating unit configured to analyze a link structure among documents on a hypertext system in which the documents are interlinked, and to calculate link distances between the documents by link transitions,
wherein the similar document specifying unit is configured to specify as a document group, documents within a link distance not more than a predetermined distance from the given document as the retrieval target, using the link distance calculating unit, and to compare the access similarities between the documents in the document group and the given document as the retrieval target with the similarity determination threshold to specify the similar document.

### Claim 4 (depends on 3)

The index generating system according to claim 3,
wherein the similarity calculating unit is configured to calculate the access similarities so as to decrease with increase in the link distance calculated by the link distance calculating unit.

### Claim 5 (depends on 2)

The index generating system according to claim 2,
wherein the access history storage unit is configured to store as the history information the frequency information including access durations corresponding to the respective access frequencies, and
wherein the similarity calculating unit is configured to calculate the access similarities so as to increase or decrease the access frequencies in the frequency information, based on the access durations.

### Claim 6 (depends on 1)

The index generating system according to claim 1,
wherein the access history storage unit is configured to store as the history information, time-series access information wherein access situations to the respective documents by the respective users are correlated with identification information of the users and access times,
wherein the similarity calculating unit is configured to calculate the access similarities, based on the access information stored in the access history storage unit, and
wherein the similar document specifying unit is configured to compare the access similarities with a similarity determination threshold to specify the similar document.

### Claim 7 (depends on 6)

The index generating system according to claim 6,
wherein the access history storage unit is configured to store as the history information the access information including access durations corresponding to the respective access situations, and
wherein the similarity calculating unit is configured to calculate the access similarities so as to increase or decrease the access situations in the access information, based on the access durations.

### Claim 8 (depends on 2)

The index generating system according to claim 2,
wherein the similar document specifying unit has a plurality of similarity determination thresholds as mentioned and is configured to specify the similar document in a plurality of patterns according to the plurality of similarity determination thresholds, and
wherein the retrieval index generating unit is configured to generate retrieval indexes in plural patterns for the given document as the retrieval target from words appearing in each of document sets specified according to the similar documents consisting of the plurality of patterns.

### Claim 9 (depends on 1)

An information retrieval system for document retrieval, comprising:
an information retrieval unit configured to retrieve a fit document fitting a retrieval request, using the retrieval index generated by the index generating system defined in claim 1.

### Claim 10 (depends on 8)

An information retrieval system for document retrieval, comprising:
an information retrieval unit configured to retrieve a fit document fitting a retrieval request, using the retrieval indexes generated by the index generating system defined in claim 8.

### Claim 11 (depends on 8, 10)

The information retrieval system according to claim 10,
wherein the information retrieval unit is configured to retrieve the fit document fitting the retrieval request, using the retrieval index of one pattern out of the plurality of patterns generated by the index generating system defined in claim 8, and to select and use the retrieval index of another pattern out of the plurality of patterns according to a number of fit documents retrieved.

### Claim 12 (depends on 9)

The information retrieval system according to claim 9, further comprising a retrieval result reducing unit configured so that if the information retrieval unit retrieves a plurality of documents in an identical site as the fit document the retrieval result reducing unit performs such reduction as to give a higher priority to a document with a larger similarity to the retrieval request.

### Claim 13 (depends on 9)

The information retrieval system according to claim 9, further comprising a retrieval result reducing unit configured so that if the information retrieval unit retrieves a plurality of documents in an identical site as the fit document the retrieval result reducing unit performs such reduction as to give a higher priority to a document with a larger number of words constituting the retrieval index.

### Claim 14 (depends on 9)

The information retrieval system according to claim 9, further comprising a retrieval result reducing unit configured so that if the information retrieval unit retrieves a plurality of documents in an identical site as the fit document the retrieval result reducing unit performs such reduction as to give a higher priority to a document with a document path thereof at a higher level out of the plurality of documents.

### Claim 15 (depends on 9)

The information retrieval system according to claim 9, further comprising a retrieval result reducing unit configured so that if the information retrieval unit retrieves a plurality of documents in an identical site as the fit document the retrieval result reducing unit performs such reduction as to give a higher priority to a document with a higher document access count, based on the history information stored in the access history storage unit.

### Claim 16 (depends on 9)

The information retrieval system according to claim 9, further comprising a retrieval result reducing unit configured so that if the information retrieval unit retrieves a plurality of documents in an identical site as the fit document the retrieval result reducing unit performs such reduction as to give a higher priority to an access start document in each session by the users, based on the history information stored in the access history storage unit.

### Claim 17

An index generating method for generating a retrieval index for document retrieval, comprising:
an access history storage step of storing history information indicating access histories to respective documents by users;
a similarity calculating step of calculating access similarities indicating similarities of the access histories between the documents, based on the history information stored in the access history storage step;
a similar document specifying step of specifying a similar document similar to a given document as a retrieval target, based on the access similarities; and
a retrieval index generating step of generating a retrieval index for the given document as the retrieval target from words appearing in a document set consisting of the given document as the retrieval target and the similar document.
