Information retrieval system, method, and program

Patent metadata

Field	Value
Publication number	US-9811584-B2
Application number	US-201214235151-A
Country	US
Kind code	B2
Filing date	May 1, 2012
Priority date	Aug 5, 2011
Publication date	Nov 7, 2017
Grant date	Nov 7, 2017

Abstract

An information retrieval method for retrieving documents relevant to a user's search inquiry from documents stored in the storage device using an arithmetic operation. The method includes receiving from a user a search query including a plurality of keywords, calculating relevance to a plurality of documents on the basis of the plurality of keywords and the influence set for each keyword, and displaying on the display device documents in the order of relevance. The method also includes displaying on the display device the influence set for each keyword, receiving changes to the displayed influence by the user, and recalculating the relevance on the basis of a change to the influence and displaying on the display device documents in the order of relevance.

First claim

The invention claimed is:

1. An information retrieval system having a storage device, a display device and a processor, wherein the processor is configured to:
receive from a user a search query including a plurality of keyword;
calculate a relevance to a plurality of documents on the basis of the plurality of keywords and an influence set for each keyword (t), wherein the relevance is calculated based on:
a coefficient (coord(q,d)) determined by a number of keywords in a search query (q) in a document (d);
a coefficient tf(t,d) determined by a frequency of a keyword (t) appearing in the document (d);
a coefficient idf(t) determined by a reciprocal of a proportion of documents containing the keyword t;
a weight (boost(t)) of the keyword t;
a coefficient indicative of weight (norm(t,d)) indicative of when a search index was created;
a coefficient (date(d)) determined by a date of the document d; and
a weight for the date (dateBoost); and
wherein the relevance (score) is calculated by:
Score( q,d )=coord( q,d )×queryNorm( q )×(Σ tf ( t,d )×( idf ( t ))2×boost( t )×norm( t,d )+(date ( d ))×(dateBoost));

display on the display device documents in the order of relevance;
display on the display device the influence set for each keyword;
receive changes to the displayed influence set from the user; and
recalculate the relevance on the basis of a change to the influence set and display on the display device documents in the order of relevance.

2. The information retrieval system of claim 1 , wherein the relevance of a document is calculated on the basis of a plurality of keywords, the influence set for each keyword, and a date.

3. The information retrieval system of claim 2 , wherein the date has the same influence as a keyword, and the influence can be changed in the same manner as a keyword.

4. The information retrieval system of claim 1 , wherein the processor is further configured to display on the display device the influence of each keyword in each retrieved document.

5. The information retrieval system of claim 1 , wherein the influence of each keyword is displayed linearly and the influence of each keyword changes in accordance with movement of a boundary of each linearly displayed keyword.

6. The information retrieval system of claim 5 , wherein the information retrieval system further comprises a pointing device, a boundary of a keyword is displayed as a slide bar on the display device, and the influence is changed by operating the slide bar using the pointing device.

7. A computer-implemented information retrieval method-comprising:
receiving, via a processor, a search query from a user, wherein the search query comprises a plurality of keywords;
calculating, via the processor, a relevance to a plurality of documents, wherein the relevance is based on the plurality of keywords and an influence set for each keyword, wherein the relevance is calculated based on a mathematical product of:
a number of keywords in a search query in a document;
a frequency of a keyword appearing in the document;
a proportion of documents containing the keyword;
a weight of the keyword; and
a weight for a date indicative of when a search index was created;

displaying on the display device, via the processor, documents in the an order of relevance;
displaying on the display device, via the processor, the influence set for each keyword;
receiving, via the processor, changes to the displayed influence by the user; and
recalculating, via the processor, the relevance on the basis of a change to the influence and displaying on the display device documents in the order of relevance.

8. The information retrieval method of claim 7 , wherein the relevance of a document is calculated on the basis of a plurality of keywords, the influence set for each keyword, and a date.

9. The information retrieval method of claim 8 , wherein the date has the same influence as a keyword, and the influence can be changed in the same manner as a keyword.

10. The information retrieval method of claim 7 , further comprising displaying on the display device the influence of each keyword in each retrieved document.

11. The information retrieval method of claim 7 , further comprising displaying the influence of each keyword linearly, and changing the influence of each keyword in accordance with movement of a boundary of each linearly displayed keyword.

12. The information retrieval method of claim 11 , wherein the information retrieval method further comprises a pointing device, a boundary of a keyword is displayed as a slide bar on the display device, and the influence is changed by operating the slide bar using the pointing device.

13. A non-transitory computer readable storage medium having computer readable program code embodied therewith, the computer readable program code executable to cause a processor to perform a method for operating an information retrieval system, the method comprising:
receiving, via a processor, a search query from a user, wherein the search query comprises a plurality of keywords;
calculating, via the processor, relevance to a plurality of documents on the basis of the plurality of keywords and an influence set for each keyword, wherein the relevance is calculated a mathematical product of:
a number of keywords in a search query in a document;
a frequency of a keyword appearing in the document;
a proportion of documents containing the keyword;
a weight of the keyword; and
a weight for a date indicative of when a search index was created;

displaying, via the processor, on a display device documents in the order of relevance;
displaying on the display device, via the processor, an influence set for each keyword;
receiving, via the processor, changes to the displayed influence by the user; and
recalculating the relevance, via the processor, on the basis of a change to the influence and displaying on the display device documents in the order of relevance.

14. The computer program product of claim 13 , wherein the relevance of a document is calculated on the basis of a plurality of keywords, the influence set for each keyword, and a date.

15. The computer program product of claim 14 , wherein the date has the same influence as a keyword, and the influence can be changed in the same manner as a keyword.

16. The computer program product of claim 13 , further comprising displaying on the display device the influence of each keyword in each retrieved document.

17. The computer program product of claim 13 , further comprising displaying the influence of each keyword linearly, and changing the influence of each keyword in accordance with movement of a boundary of each linearly displayed keyword.

18. The computer program product of claim 17 , wherein the information retrieval method further comprises a pointing device, a boundary of a keyword is displayed as a slide bar on the display device, and the influence is changed by operating the slide bar using the pointing device.

Assignees

• Adachi Noritaka
• Kurokawa Hiroshi
• Matsuoka Kensuke
• Murakami Yosuke
• IBM

Inventors

• Adachi Noritaka
• Kurokawa Hiroshi
• Matsuoka Kensuke
• Murakami Yosuke

Classifications

• G06F17/30675

  Physics · mapped topic

• G06F17/30386

  Physics · mapped topic

• G06F17/30648 Primary

  Physics · mapped topic

• G06F16/334 Primary

  Query execution (filtering based on additional data G06F16/335) · CPC title

• G06F16/3326 Primary

  using relevance feedback from the user, e.g. relevance feedback on documents, documents sets, document terms or passages · CPC title