Date ambiguity resolution

• US 8,239,350 B1
• Filed: 05/08/2007
• Issued: 08/07/2012
• Est. Priority Date: 05/08/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Vespe, David, Hogue, Andrew W.
• Primary Examiner(s): Pulliam, Christyann
• Application Number:
  US11/745,605
• Time in Patent Office: 1,918 Days
• Field of Search: None
• US Class Current: 707/687
• CPC Class Codes: G06F 40/126 Character encoding

Abstract

A system and method for resolving ambiguities in date values associated with an attribute in a memory of the computer system. If a first text string conforms to one or more date formats, a confidence value is assigned for each of the date formats for the first text string based on the amount of specificity with which the first text string conforms to each date format. Similarly, if a second text string conforms to one or more date formats, a confidence value is assigned for each of the date formats for the second text string based on the amount of specificity with which the second text string conforms to each date format. The date format with the highest confidence value for the first text string and the date format with the highest confidence value for the second text string are merged to obtain a date value for the attribute.

Claims

1. A computer-implemented method for resolving ambiguities in date values associated with an attribute of an entity, the method comprising:

at a computer system including one or more processors and memory storing one or more programs, the one or more processors executing the one or more programs to perform the operations of;

obtaining a first text string associated with an attribute of an entity, wherein the first text string is extracted from a first web document;

determining that the first text string conforms to one or more date formats;

assigning a first confidence value for each of the date formats for the first text string based on a first number of unknown variables that remain when interpreting the first text string using each of the date formats;

obtaining a second text string associated with the attribute of the entity, wherein the second text string is extracted from a second web document;

determining that the second text string conforms to one or more of the date formats;

assigning a second confidence value for each of the date formats for the second text string based on a second number of unknown variables that remain when interpreting the second text string using each of the date formats;

determining a first date string expressed in a date format with a highest first confidence value for the first text string;

determining a second date string expressed in a date format with a highest second confidence value for the second text string; and

merging a first subset of the first date string and a second subset of the second date string to obtain a date value for the attribute.

2. The method of claim 1, further comprising:

determining whether the obtained date value is consistent with rules related to the attribute.

3. The method of claim 1, further comprising:

determining whether the first text string is consistent with rules related to the attribute.

4. The method of claim 1, further comprising:

applying a date format for the obtained date value to all text strings that conform to one or more date formats from at least one domain from which the obtained date value was obtained.

5. The method of claim 1, wherein the first text string is interpreted as a range of time.

6. The method of claim 1, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that do not conflict.

7. The method of claim 1, wherein the merging of the date formats retains component;

of the date format for the first text string and components of the date format for the second text string that are common to both formats.

8. The method of claim 1, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are not ambiguous.

9. The method of claim 1, wherein a respective text string is obtained from a respective date source, and wherein the method further comprises:

repeating the operations of claim 1 until a threshold number of date sources with identical date formats for the attribute are obtained.

10. The method of claim 1, wherein the attribute is a date of birth.

11. The method of claim 1, wherein the steps are repeated for multiple attributes of one entity.

12. The method of claim 1, wherein one of the date formats is a United States date format.

13. The method of claim 1, wherein one of the date formats is a European date format.

14. The method of claim 1, wherein the confidence value for a respective date format is further based on an amount of specificity of the respective date format, the amount of specificity of the respective date format being selected from the group consisting of:

a number of unknown variables in the respective date format;

a precision of the respective date format.

15. The method of claim 1, further comprising:

determining a threshold number of date formats to be merged to obtain a date value.

25. The system of claim 16, wherein a respective text string is obtained from a respective date source, and wherein the system further comprises instructions to:

repeat the instructions of claim 14 until a threshold number of date sources with identical date formats for the attribute are obtained.

16. A computer system for resolving ambiguities in date values associated with an attribute of an entity, the computer system comprising:

one or more processors;

memory; and

one or more programs stored in the memory, the one or more programs comprising instructions to;

obtain a first text string associated with an attribute of an entity, wherein the first text string is extracted from a first web document;

determine if that the first text string conforms to one or more date formats;

assign a first confidence value for each of the date formats for the first text string based on a first number of unknown variables that remain when interpreting the first text string using each of the date formats;

obtain a second text string associated with the attribute of the entity, wherein the second text string is extracted from a second web document;

determine if that the second text string conforms to one or more date formats;

assign a second confidence value for each of the date formats for the second text string based on a second number of unknown variables that remain when interpreting the second text string using each of the date formats;

determine a first date string expressed in a date format with a highest first confidence value for the first text string;

determine a second date string expressed in a date format with a highest second confidence value for the second text string; and

merge a first subset of the first date string and a second subset of the second date string to obtain a date value for the attribute.

17. The system of claim 16, further comprising instructions to:

determine whether the obtained date value is consistent with rules related to the attribute.

18. The system of claim 16, further comprising instructions to:

determine whether the first text string is consistent with rules related to the attribute.

19. The computer system of claim 16, wherein the confidence value for a respective date format is further based on an amount of specificity of the respective date format, the amount of specificity of the respective date format being selected from the group consisting of:

a number of unknown variables of in the respective date format;

a precision of the respective date format.

20. The system of claim 16, further comprising instructions to:

apply a date format for the obtained date value to all text strings that conform to one or more date formats from at least one domain from which the obtained date value was obtained.

21. The system of claim 16, wherein the first text string is interpreted as a range of time.

22. The system of claim 16, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that do not conflict.

23. The system of claim 16, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are common to both formats.

24. The system of claim 16, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are not ambiguous.

26. The system of claim 16, wherein the attribute is a date of birth.

27. The system of claim 16, wherein the instructions are repeated for multiple attributes of one entity.

28. The system of claim 16, wherein one of the date formats is a United States date format.

29. The system of claim 16, wherein one of the date formats is a European date format.

30. The system of claim 16, further comprising instructions to:

determine a threshold number of date formats to be merged to obtain a date value.

34. The computer program product of claim 19, the computer-readable storage medium further comprising:

program code for determining a threshold number of date formats to be merged to obtain a date value.

52. The method of claim 46, wherein a respective text string is obtained from a respective date source, and wherein the method further comprises:

repeating the operations of claim 24 until a threshold number of date sources with identical date formats for the attribute are obtained.

31. A computer program product comprising a computer-readable storage medium, for resolving ambiguities in date values associated with an attribute of an entity, the computer-readable storage medium comprising:

program code for obtaining a first text string associated with an attribute of an entity, wherein the first text string is extracted from a first web document;

program code for determining that the first text string conforms to one or more date formats;

program code for assigning a first confidence value for each of the date formats for the first text string based on a first number of unknown variables that remain when interpreting the first text string using each of the date formats;

program code for obtaining a second text string associated with the attribute of the entity, wherein the second text string is extracted from a second web document;

program code for determining if that the second text string conforms to one or more of the date formats;

program code for assigning a second confidence value for each of the date formats for the second text string based on a second number of unknown variables that remain when interpreting the second text string using each of the date formats;

program code determining a first date string expressed in a date format with a highest first confidence value for the first text string;

program code determining a second date string expressed in a date format with a highest second confidence value for the second text string; and

program code for merging a first subset of the first date string and a second subset of the second date string to obtain a date value for the attribute.

32. The computer program product of claim 31, the computer-readable storage medium further comprising:

program code for determining whether the obtained date value is consistent with rules related to the attribute.

33. The computer program product of claim 31, the computer-readable storage medium further comprising:

program code for determining whether the first text string is consistent with rules related to the attribute.

35. The computer program product of claim 31, wherein the confidence value for a respective date format is further based on an amount of specificity of the respective date format, the amount of specificity of the respective date format being selected from the group consisting of:

a number of unknown variables ef-in the respective date format;

a precision of the respective date format.

36. The computer program product of claim 31, the computer-readable storage medium further comprising:

program code for applying a date format for the obtained date value to all text strings that conform to one or more date formats from at least one domain from which the obtained date value was obtained.

37. The computer program product of claim 31, wherein the first text string is interpreted as a range of time.

38. The computer program product of claim 31, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that do not conflict.

39. The computer program product of claim 31, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are common to both formats.

40. The computer program product of claim 31, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are not ambiguous.

41. The computer program product of claim 31, wherein a respective text string is obtained from a respective date source, and wherein the computer program product further comprises:

program code for repeating the program codes of claim 31 until a threshold number of date sources with identical date formats for the attribute are obtained.

42. The computer program product of claim 31, wherein the attribute is a date of birth.

43. The computer program product of claim 31, wherein the program codes are repeated for multiple attributes of one entity.

44. The computer program product of claim 31, wherein one of the date formats is a United States date format.

45. The computer program product of claim 31, wherein one of the date formats is a European date format.

46. A computer-implemented method for resolving ambiguities in date values associated with an attribute of an entity, the method comprising:

identifying a plurality of web documents associated with an attribute of an entity;

for each web document in the plurality of web documents,

obtaining, from the web document, one or more text strings associated with the attribute of the entity;

identifying one or more date formats for at least one of the one or more text strings; and

assigning confidence values to each of the one or more date formats for the at least one of the one or more text strings based on a number of unknown variables that remain when interpreting the at least one of the one or more text strings using each of the one or more date formats;

determining date strings expressed in date formats with highest confidence values for the at least one of the one or more text strings; and

merging subsets of the date strings to obtain a date value for the attribute.

47. The method of claim 46, wherein the confidence value for a respective date format is further based on an amount of specificity of the respective date format, the amount of specificity of the respective date format being selected from the group consisting of:

a number of unknown variables in the respective date format;

a precision of the respective date format.

48. The method of claim 46, further comprising:

applying a date format for the obtained date value to all text strings that conform to one or more date formats from at least one domain from which the obtained date value was obtained.

49. The method of claim 46, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that do not conflict.

50. The method of claim 46, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are common to both formats.

51. The method of claim 46, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are not ambiguous.

53. A computer system for resolving ambiguities in date values associated with an attribute of an entity, the computer system comprising:

one or more processors;

memory; and

one or more programs stored in the memory, the one or more programs comprising instructions to;

identify a plurality of web documents associated with an attribute of an entity;

for each web document in the plurality of web documents,

obtain, from the web document, at least two text strings associated with the attribute of the entity;

identify one or more date formats for at least two text strings; and

assign confidence values to each of the one or more date formats for the at least two text strings based on a number of unknown variables that remain when interpreting the at least one of the one or more at least two text strings using each of the one or more date formats;

determine date strings expressed in date formats with highest confidence values for the at least two text strings; and

merge subsets of the date strings to obtain a date value for the attribute.

54. The system of claim 53, wherein the confidence value for a respective date format is further based on an amount of specificity of the respective date format, the amount of specificity of the respective date format being selected from the group consisting of:

a number of unknown variables in the respective date format;

a precision of the respective date format.

55. The system of claim 53, further comprising instructions to:

apply a date format for the obtained date value to all text strings that conform to one or more date formats from at least one domain from which the obtained date value was obtained.

56. The system of claim 53, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that do not conflict.

57. The system of claim 53, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are common to both formats.

58. The system of claim 53, wherein the merging of the date formats retains components of the date format for the first text string and components of the date format for the second text string that are not ambiguous.

59. The system of claim 53, wherein a respective text string is obtained from a respective date source, and wherein the system further comprises instructions to:

repeat the instructions of claim 53 until a threshold number of date sources with identical date formats for the attribute are obtained.