Concept bridge and method of operating the same

• US 8,312,034 B2
• Filed: 06/21/2006
• Issued: 11/13/2012
• Est. Priority Date: 06/24/2005
• Status: Active Grant

• Current Assignee: Brainspace Corporation (Reveal Data Corp.)
• Original Assignee: PureDiscovery Corporation (Reveal Data Corp.)
• Inventors: Hagar, David Adam, Jernigan, Stephen Scott, Copps, David Seigert
• Primary Examiner(s): Vu, Bai D.
• Application Number:
  US11/472,026
• Publication Number:
  US 20070011151A1
• Time in Patent Office: 2,337 Days
• Field of Search: None
• US Class Current: 707/765
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

A concept bridge employable with a search engine, method of operating the same and computer information system employing the concept bridge and method. In one embodiment, the concept bridge includes an extractor configured to derive concept terms by extracting significant terms from search text and inferring relevant terms therefrom. The concept bridge also includes a query generator configured to generate a query consistent with an index of a search engine as a function of the concept terms.

Claims

1. A computer system comprising:

a processor;

a computer readable memory encoded with logic that when executed by the processor is operable to perform operations comprising;

deriving concept terms by extracting significant terms from search text and inferring relevant terms from the extracted significant terms using a concept matrix, the concept matrix derived from a plurality of sample documents using singular value decomposition of a term-document matrix, the concept matrix identifying a latent pattern of word usage in the plurality of sample documents around a concept, the concept matrix further providing a term-term similarity between the search text and terms in the plurality of sample documents; and

generating a query for a search engine, the query comprising a search expression having at least one of the derived concept terms.

2. The computer system of claim 1 further comprising a filter configured to remove noise terms that are unrelated to said significant terms.

3. The computer system of claim 1 wherein said search text is selected from the group consisting of:

a document,

an e-mail, and

a Web page.

4. The computer system of claim 1 wherein the logic is further operable to cluster concept terms in synonym groups with an operator therebetween.

5. The computer system of claim 1 wherein the logic is further operable to generate the query conforming to a Boolean expression.

6. The computer system of claim 1 wherein the computer system is employable with another search engine and the logic is further operable to generate a different query consistent with an index of said another search engine as a function of said concept terms.

7. The computer system of claim 1, wherein the concept matrix comprises a compressed matrix derived by compressing a term by document matrix using singular value decomposition, the term by document matrix derived from the plurality of sample documents using latent semantic analysis and comprising weighted term frequencies.

8. A method of generating a query for a search engine, comprising:

deriving concept terms by extracting significant terms from search text and inferring relevant terms from the extracted significant terms using a concept matrix, the concept matrix derived from a plurality of sample documents using singular value decomposition of a term-document matrix, the concept matrix identifying a latent pattern of word usage in the plurality of sample documents around a concept, the concept matrix further providing a term-term similarity between the search text and terms in the plurality of sample documents; and

generating a query for a search engine, the query comprising a search expression having at least one of the derived concept terms.

9. The method as recited in claim 8 wherein said search text is selected from the group consisting of:

a document,

an e-mail, and

a Web page.

10. The method as recited in claim 8 further comprising building said concept matrix.

11. The method as recited in claim 8 wherein said generating includes clustering concept terms in synonym groups with an operator therebetween.

12. The method as recited in claim 8 wherein said query conforms to a Boolean expression.

13. The method as recited in claim 8 wherein said method is employable with another search engine and said generating includes generating a different query consistent with an index of said another search engine as a function of said concept terms.

14. The method of claim 8, wherein the concept matrix comprises a compressed matrix derived by compressing a term by document matrix using singular value decomposition, the term by document matrix derived from the plurality of sample documents using latent semantic analysis and comprising weighted teen frequencies.

15. The method as recited in claim 8, further comprising:

transmitting the generated query to the search engine via an application program interface;

receiving from the search engine a result associated with the query; and

displaying the result on a display.

16. A computer information system, comprising:

a first search engine having a first index;

a second search engine having a second index different from said first index; and

a computer readable memory encoded with logic that when executed by a processor is operable to perform operations comprising;

deriving concept terms by extracting significant terms from search text and inferring relevant terms from the extracted significant terms using a concept matrix, the concept matrix derived from a plurality of sample documents using singular value decomposition of a term-document matrix, the concept matrix identifying a latent pattern of word usage in the plurality of sample documents around a concept, the concept matrix further providing a term-term similarity between the search text and terms in the plurality of sample documents; and

generating a first query for said first index and a second query for said second index, the first and second queries comprising search expressions having at least one of the derived concept terms.

17. The computer information system as recited in claim 16 wherein the logic is further operable to remove noise terms that are unrelated to said significant terms.

18. The computer information system as recited in claim 16 wherein said search text is selected from the group consisting of:

a document,

an e-mail, and

a Web page.

19. The computer information system as recited in claim 16 wherein the logic is further operable to cluster concept terms in synonym groups with an operator therebetween.

20. The computer information system as recited in claim 16 further comprising a ranking subsystem configured to rank search results from at least one of said first and second search engines.

21. The computer information system of claim 16, wherein the concept matrix comprises a compressed matrix derived by compressing a term by document matrix using singular value decomposition, the term by document matrix derived from the plurality of sample documents using latent semantic analysis and comprising weighted term frequencies.