System, method and computer program for facet analysis

• US 7,860,817 B2
• Filed: 09/09/2009
• Issued: 12/28/2010
• Est. Priority Date: 03/30/2005
• Status: Active Grant

• Current Assignee: Primal Fusion Inc.
• Original Assignee: Primal Fusion Inc.
• Inventors: Sweeney, Peter, Yao, Zhengrong
• Primary Examiner(s): Holmes, Michael
• Application Number:
  US12/556,349
• Publication Number:
  US 20100036790A1
• Time in Patent Office: 475 Days
• Field of Search: 706/48
• US Class Current: 706/48
• CPC Class Codes: G06F 16/84 Mapping; Conversion
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

Automated facet analysis of input information selected from a domain of information in accordance with a source data structure is described. Facet analysis may proceed by discovering at least one of facets, facet attributes, and facet attribute hierarchies of the input information using pattern augmentation and statistical analyses to identify patterns of facet attribute relationships in the input information.

Claims

1. In a computer system comprising at least one hardware processor and at least one tangible storage medium, a method of identifying elemental facet attribute relationships in a domain of information comprising a plurality of concepts, wherein the method comprises:

identifying a plurality of concept relationships between pairs of concepts in the plurality of concepts;

identifying at least one facet attribute in each of at least some of the pairs of concepts;

based, at least in part, on the identified plurality of concept relationships, identifying a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts; and

determining whether to include a candidate facet attribute relationship from the set of candidate facet attribute relationships in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the candidate facet attribute relationship.

2. The method of claim 1, wherein the pervasiveness of the candidate facet attribute relationship is measured as the likelihood that the facet attribute relationship holds true in all concept relationships in which it presents.

3. The method of claim 2, wherein the likelihood that the candidate facet attribute relationship holds true in all concept relationships in which it presents is determined based on a frequency with which the facet attributes related by the candidate facet attribute relationship present together in the domain of information.

4. The method of claim 2, wherein the plurality of concept relationships comprises a first concept relationship between a first of the plurality of concepts and a second of the plurality of concepts, and wherein the act of identifying at least one facet attribute in at least some of the pairs of concepts further comprises:

identifying first and second facet attributes in the first of the plurality of concepts and identifying third and fourth attributes in the second of the plurality of concepts.

5. The method of claim 4, wherein the act of identifying a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts further comprises:

identifying a first facet attribute relationship between the first and third facet attributes;

identifying a second facet attribute relationship between the first and fourth facet attributes;

identifying a third facet attribute relationship between the second and third facet attributes; and

identifying a fourth facet attribute relationship between the second and fourth facet attributes.

6. The method of claim 5, wherein the act of determining whether to include a candidate facet attribute relationship in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the facet attribute relationship further comprises:

determining whether to include the first facet attribute relationship between the first and third facet attributes in the modified set of candidate facet attribute relationships based on a prevalence of a relationship between the first and third facet attributes in the identified plurality of concept relationships.

7. The method of claim 1, wherein the act of identifying at least one facet attribute in each of at least some of the pairs of concepts further comprises:

identifying a plurality of keywords in the at least some of the pairs of concepts; and

identifying the plurality of facet attributes in the plurality of keywords.

8. The method of claim 7, wherein the act of identifying the plurality of facet attributes in the plurality of keywords further comprises:

defining at least one pattern for use as criteria for identifying facet attribute candidates;

applying the at least one pattern against the plurality of keywords; and

based on the application of the at least one pattern, identifying a candidate set of facet attributes.

9. The method of claim 8, further comprising:

comparing the candidate facet attributes in the candidate set of facet attributes to identify inconsistent candidate facet attributes; and

removing inconsistent candidate facet attributes from the candidate set of facet attributes.

28. The method of claim 1, wherein the method is a method for identifying elemental facet attribute relationships by stripping away noise.

10. At least one computer readable medium encoded with computer instructions that, when executed in a computer system perform a method of identifying elemental facet attribute relationships in a domain of information comprising a plurality of concepts, wherein the method comprises:

identifying a plurality of concept relationships between pairs of concepts in the plurality of concepts;

identifying at least one facet attribute in each of at least some of the pairs of concepts;

based, at least in part, on the identified plurality of concept relationships, identifying a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts; and

determining whether to include a candidate facet attribute relationship from the set of candidate facet attribute relationships in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the candidate facet attribute relationship.

11. The at least one computer readable medium of claim 10, wherein the pervasiveness of the candidate facet attribute relationship is measured as the likelihood that the facet attribute relationship holds true in all concept relationships in which it presents.

12. The at least one computer readable medium of claim 11, wherein the likelihood that the candidate facet attribute relationship holds true in all concept relationships in which it presents is determined based on a frequency with which the facet attributes related by the candidate facet attribute relationship present together in the domain of information.

13. The at least one computer readable medium of claim 11, wherein the plurality of concept relationships comprises a first concept relationship between a first of the plurality of concepts and a second of the plurality of concepts, and wherein the act of identifying at least one facet attribute in at least some of the pairs of concepts further comprises:

identifying first and second facet attributes in the first of the plurality of concepts and identifying third and fourth attributes in the second of the plurality of concepts.

14. The at least one computer readable medium of claim 13, wherein the act of identifying a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts further comprises:

identifying a first facet attribute relationship between the first and third facet attributes;

identifying a second facet attribute relationship between the first and fourth facet attributes;

identifying a third facet attribute relationship between the second and third facet attributes; and

identifying a fourth facet attribute relationship between the second and fourth facet attributes.

15. The at least one computer readable medium of claim 14, wherein the act of determining whether to include a candidate facet attribute relationship in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the facet attribute relationship further comprises:

determining whether to include the first facet attribute relationship between the first and third facet attributes in the modified set of candidate facet attribute relationships based on a prevalence of a relationship between the first and third facet attributes in the identified plurality of concept relationships.

16. The at least one computer readable medium of claim 10, wherein the act of identifying at least one facet attribute in each of at least some of the pairs of concepts further comprises:

identifying a plurality of keywords in the at least some of the pairs of concepts; and

identifying the plurality of facet attributes in the plurality of keywords.

17. The at least one computer readable medium of claim 16, wherein the act of identifying the plurality of facet attributes in the plurality of keywords further comprises:

defining at least one pattern for use as criteria for identifying facet attribute candidates;

applying the at least one pattern against the plurality of keywords; and

based on the application of the at least one pattern, identifying a candidate set of facet attributes.

18. The at least one computer readable medium of claim 17, wherein the method further comprises:

comparing the candidate facet attributes in the candidate set of facet attributes to identify inconsistent candidate facet attributes; and

removing inconsistent candidate facet attributes from the candidate set of facet attributes.

29. The at least one computer readable medium of claim 10, wherein the computer instructions, when executed in the computer system, perform a method of identifying elemental facet attribute relationships by stripping away noise.

19. A computer comprising:

at least one tangible memory that stores processor-executable instructions for performing a method identifying elemental facet attribute relationships in a domain of information comprising a plurality of concepts; and

at least one hardware processor, coupled to the at least one tangible memory, configured to execute the processor-executable instructions to;

identify a plurality of concept relationships between pairs of concepts in the plurality of concepts;

identify at least one facet attribute in each of at least some of the pairs of concepts;

based, at least in part, on the identified plurality of concept relationships, identify a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts; and

determine whether to include a candidate facet attribute relationship from the set of candidate facet attribute relationships in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the candidate facet attribute relationship.

20. The computer of claim 19, wherein the pervasiveness of the candidate facet attribute relationship is measured as the likelihood that the facet attribute relationship holds true in all concept relationships in which it presents.

21. The computer of claim 20, wherein the likelihood that the candidate facet attribute relationship holds true in all concept relationships in which it presents is determined based on a frequency with which the facet attributes related by the candidate facet attribute relationship present together in the domain of information.

22. The computer of claim 20, wherein the plurality of concept relationships comprises a first concept relationship between a first of the plurality of concepts and a second of the plurality of concepts, and wherein the at least one hardware processor executes the processor executable instructions to identify the at least one facet attribute in the at least some of the pairs of concepts by identifying first and second facet attributes in the first of the plurality of concepts and identifying third and fourth attributes in the second of the plurality of concepts.

25. The computer of claim 19, wherein the at least one hardware processor executes the processor executable instructions to identify at least one facet attribute in each of at least some of the pairs of concepts by:

identifying a plurality of keywords in the at least some of the pairs of concepts; and

identifying the plurality of facet attributes in the plurality of keywords.

26. The computer of claim 25, wherein the at least one hardware processor executes the processor executable instructions to identify the plurality of facet attributes in the plurality of keywords by:

defining at least one pattern for use as criteria for identifying facet attribute candidates;

applying the at least one pattern against the plurality of keywords; and

based on the application of the at least one pattern, identifying a candidate set of facet attributes.

27. The computer of claim 26, wherein the at least one hardware processor executes the processor executable instructions to:

comparing the candidate facet attributes in the candidate set of facet attributes to identify inconsistent candidate facet attributes; and

removing inconsistent candidate facet attributes from the candidate set of facet attributes.

30. The computer of claim 19, wherein the processor-executable instructions are instructions for performing a method identifying elemental facet attribute relationships by stripping away noise.

23. The computer of 22, wherein the at least one hardware processor executes the processor executable instructions to identify a set of candidate facet attribute relationships between the facet attributes in the at least some of the pairs of concepts by:

identifying a first facet attribute relationship between the first and third facet attributes;

identifying a second facet attribute relationship between the first and fourth facet attributes;

identifying a third facet attribute relationship between the second and third facet attributes; and

identifying a fourth facet attribute relationship between the second and fourth facet attributes.

24. The computer claim 23, wherein the at least one hardware processor executes the processor executable instructions to determine whether to include a candidate facet attribute relationship in a modified set of candidate facet attribute relationships based, at least in part, on the pervasiveness of the facet attribute relationship by:

determining whether to include the first facet attribute relationship between the first and third facet attributes in the modified set of candidate facet attribute relationships based on a prevalence of a relationship between the first and third facet attributes in the identified plurality of concept relationships.