Information searching method of profile information, program, recording medium, and apparatus

• US 6,915,295 B2
• Filed: 03/22/2002
• Issued: 07/05/2005
• Est. Priority Date: 12/13/2001
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Sato, Akira, Ando, Takahisa, Okamoto, Seishi, Inakoshi, Hiroya
• Primary Examiner(s): Pardo, Thuy N.
• Application Number:
  US10/102,773
• Publication Number:
  US 20030115193A1
• Time in Patent Office: 1,201 Days
• Field of Search: 707/3, 707/4, 707/5, 707/6, 707/10, 707/1, 707/104, 705/8, 705/10, 345/810, 434/322
• US Class Current: 707/728
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: Y10S 707/947 Human resources
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

In a candidate data searching step, a plurality of similar candidate data is searched from a profile data group having a peculiar attribute value every different attribute on the basis of an input in which one or a plurality of attribute values have been designated. In a significance calculating step, significance regarding the attribute is calculated on the basis of the searched candidate data and non-candidate data. In a score calculating step, a score of each candidate data is calculated on the basis of the calculated significance. In a search result outputting step, each candidate data is ranked on the basis of the calculated score, and the ranked upper candidate data is outputted as a search result.

Claims

1. An information searching method of profile information, comprising:

candidate data searching wherein on the basis of input data in which attribute values have been designated, a plurality of similar candidate data is searched from a profile data group having a peculiar attribute value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

search result outputting wherein each candidate data is ranked on the basis of said calculated score and said ranked upper candidate data is outputted as a search result.

2. A method according to claim 1, wherein

in said candidate data searching the score is obtained with respect to each profile data and the candidate data is searched, and

in said score calculating the score obtained by said candidate data searching is added to the score calculated on the basis of said significance, thereby calculating the score of each of said candidate data.

3. A method according to claim 1, wherein in said significance calculating the significance of every different attribute is calculated on the basis of said candidate data and said non-candidate data.

4. A method according to claim 3, wherein in said significance calculating an information gain is calculated as said significance of every different attribute.

5. A method according to claim 4, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said information gain as a weight.

6. A method according to claim 1, wherein in said significance calculating the significance of the different attribute values included in the same attribute is calculated.

7. A method according to claim 6, wherein in said significance calculating an MVDM (Modified Value Difference Metric) is calculated as said significance of the different attribute values.

8. A method according to claim 7, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said MVDM (Modified Value Difference Metric) as a weight.

9. A method according to claim 1, wherein

in said score calculating the score of said non-candidate data is further calculated on the basis of said calculated significance, and

in said search result outputting each of said candidate data and said non-candidate data is ranked on the basis of said calculated score and said ranked upper candidate data is outputted as a search result.

10. A method according to claim 1, wherein in said search result outputting a predetermined number of said ranked upper candidate data or the candidate data having the scores which are equal to or larger than a predetermined threshold value is outputted as said search result.

11. A method according to claim 1, wherein in said search result outputting said calculated significance is added to the attribute value of profile data which is outputted as said search result and the resultant attribute value is outputted.

12. A method according to claim 1, wherein in said search result outputting if an output per viewpoint has been set, the attribute having the large significance among said calculated significance is selected as a viewpoint, each of said candidate data is sorted every different attribute value included in said selected attribute of the viewpoint, each of said candidate data sorted every attribute value is ranked on the basis of said calculated score, and said ranked upper candidate data is outputted as a search result per viewpoint.

13. An information searching method of profile information, comprising:

candidate data searching wherein on the basis of input data in which attribute values including an uninput attribute value have been designated, a plurality of similar candidate data is searched from a profile data group including the attribute values having a peculiar value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

presumption result outputting wherein similar data to said input attribute value is determined from said candidate data in accordance with the score of each of said candidate data, and the uninput attribute value of said input data is presumed on the basis of distribution of the uninput attribute values in said similar data and outputted.

14. A method according to claim 13, wherein

in said candidate data searching the score is obtained with respect to each profile data and the candidate data is searched, and

in said score calculating the score obtained by said candidate data searching is added to the score calculated on the basis of said significance, thereby calculating the score of each of said candidate data.

15. A method according to claim 13, wherein in said significance calculating the significance of every different attribute is calculated on the basis of said candidate data and said non-candidate data.

16. A method according to claim 15, wherein in said significance calculating an information gain is calculated as said significance of every different attribute.

17. A method according to claim 16, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said information gain as a weight.

18. A method according to claim 13, wherein in said significance calculating the significance of the different attribute values included in the same attribute is calculated.

19. A method according to claim 18, wherein in said significance calculating an MVDM (Modified Value Difference Metric) is calculated as said significance of the different attribute values.

20. A method according to claim 19, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said MVDM (Modified Value Difference Metric) as a weight.

21. A method according to claim 13, wherein in said presumption result outputting reliability calculated on the basis of the score of each of said candidate data is added to said presumed uninput attribute value and the resultant attribute value is outputted.

22. A method according to claim 21, wherein in said presumption result outputting said reliability is equal to a value obtained by dividing the score which enables the presumption value of each of said candidate data by a sum of the scores.

23. A program for searching profile information, wherein said program allows a computer to execute:

candidate data searching wherein on the basis of input data in which attribute values have been designated, a plurality of similar candidate data is searched from a profile data group having a peculiar attribute value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

search result outputting wherein each candidate data is ranked on the basis of said calculated score and said ranked upper candidate data is outputted as a search result.

24. A program according to claim 23, wherein

in said candidate data searching the score is obtained with respect to each profile data and the candidate data is searched, and

in said score calculating the score obtained by said candidate data searching is added to the score calculated on the basis of said significance, thereby calculating the score of each of said candidate data.

25. A program according to claim 23, wherein in said significance calculating the significance of every different attribute is calculated on the basis of said candidate data and said non-candidate data.

26. A program according to claim 25, wherein in said significance calculating an information gain is calculated as said significance of every different attribute.

27. A program according to claim 26, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said information gain as a weight.

28. A program according to claim 23, wherein in said significance calculating the significance of the different attribute values included in the same attribute is calculated.

29. A program according to claim 28, wherein in said significance calculating an MVDM (Modified Value Difference Metric) is calculated as said significance of the different attribute values.

30. A program according to claim 29, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said MVDM (Modified Value Difference Metric) as a weight.

31. A program according to claim 23, wherein

in said score calculating the score of said non-candidate data is further calculated on the basis of said calculated significance, and

in said search result outputting each of said candidate data and said non-candidate data is ranked on the basis of said calculated score and said ranked upper candidate data is outputted as a search result.

32. A program according to claim 23, wherein in said search result outputting a predetermined number of said ranked upper candidate data or the candidate data having the scores which are equal to or larger than a predetermined threshold value is outputted as said search result.

33. A program according to claim 23, wherein in said search result outputting said calculated significance is added to the attribute value of profile data which is outputted as said search result and the resultant attribute value is outputted.

34. A program according to claim 23, wherein in said search result outputting if an output per viewpoint has been designated, the attribute having the large significance among said calculated significance is selected as a viewpoint, each of said candidate data is sorted every different attribute value included in said selected attribute of the viewpoint, each of said candidate data sorted every attribute value is ranked on the basis of said calculated score, and said ranked upper candidate data is outputted as a search result per viewpoint.

35. A program for searching profile information, wherein said program allows a computer to execute:

candidate data searching wherein on the basis of input data in which attribute values including an uninput attribute value have been designated, a plurality of similar candidate data is searched from a profile data group including the attribute values having a peculiar value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

presumption result outputting wherein similar data to said input attribute value is determined from said candidate data in accordance with the score of each of said candidate data, and the uninput attribute value of said input data is presumed on the basis of distribution of the uninput attribute values in said similar data and outputted.

36. A program according to claim 35, wherein

in said candidate data searching the score is obtained with respect to each profile data and the candidate data is searched, and

in said score calculating the score obtained by said candidate data searching is added to the score calculated on the basis of said significance, thereby calculating the score of each of said candidate data.

37. A program according to claim 35, wherein in said significance calculating the significance of every different attribute is calculated on the basis of said candidate data and said non-candidate data.

38. A program according to claim 37, wherein in said significance calculating an information gain is calculated as said significance of every different attribute.

39. A program according to claim 38, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said information gain as a weight.

40. A program according to claim 35, wherein in said significance calculating the significance of the different attribute values included in the same attribute is calculated.

41. A program according to claim 40, wherein in said significance calculating an MVDM (Modified Value Difference Metric) is calculated as said significance of the different attribute values.

42. A program according to claim 41, wherein in said score calculating similarity is calculated as said score of each of said candidate data by using said MVDM (Modified Value Difference Metric) as a weight.

43. A program according to claim 35, wherein in said presumption result outputting reliability calculated on the basis of the score of each of said candidate data is added to said presumed uninput attribute value and the resultant attribute value is outputted.

44. A program according to claim 43, wherein in said presumption result outputting said reliability is equal to a value obtained by dividing the score which enables the presumption value of each of said candidate data by a sum of the scores.

45. A computer-readable recording medium which stores a program for allowing a computer to execute:

candidate data searching wherein on the basis of input data in which attribute values have been designated, a plurality of similar candidate data is searched from a profile data group having a peculiar attribute value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

search result outputting wherein each candidate data is ranked on the basis of said calculated score and said ranked upper candidate data is outputted as a search result.

46. A computer-readable recording medium which stores a program for allowing a computer to execute:

candidate data searching wherein on the basis of input data in which attribute values including an uninput attribute value have been designated, a plurality of similar candidate data is searched from a profile data group including the attribute values having a peculiar attribute value every different attribute;

significance calculating wherein significance regarding the attribute is calculated on the basis of said searched candidate data and non-candidate data;

score calculating wherein a score of each of said candidate data is calculated on the basis of said calculated significance; and

presumption result outputting wherein similar data to said input attribute value is determined from said candidate data in accordance with the score of each of said candidate data, and the uninput attribute value of said input data is presumed on the basis of distribution of the uninput attribute values in said similar data and outputted.

47. An information searching apparatus comprising:

a data storing unit in which a plurality of profile data each having a peculiar attribute value every different attribute has been stored;

a candidate data searching unit which searches a plurality of similar candidate data from the profile data group stored in said data storing unit on the basis of input data in which attribute values have been designated;

a significance calculating unit which calculates significance regarding the attribute on the basis of said searched candidate data and non-candidate data;

a score calculating unit which calculates a score of each of said candidate data on the basis of said calculated significance; and

a search result output unit which ranks each of said candidate data on the basis of said calculated score and outputs said ranked upper candidate data as a search result.

48. An information searching apparatus comprising:

a data storing unit in which a plurality of profile data each including attribute values having a peculiar attribute value every different attribute has been stored;

a candidate data searching unit which searches a plurality of similar candidate data from the profile data group stored in said data storing unit on the basis of input data in which attribute values including an uninput attribute value have been designated;

a significance calculating unit which calculates significance regarding the attribute on the basis of said searched candidate data and non-candidate data;

a score calculating unit which calculates a score of each of said candidate data on the basis of said calculated significance; and

a presumption result output unit which determines similar data to said input attribute value from said candidate data in accordance with the score of each of said candidate data, presumes the uninput attribute value of said input data on the basis of distribution of the uninput attribute values in said similar data, and outputs the presumed uninput attribute value.