System and method for determining quality of written product reviews in an automated manner

• US 20050034071A1
• Filed: 08/08/2003
• Published: 02/10/2005
• Est. Priority Date: 08/08/2003
• Status: Active Grant

• Current Assignee: CBS Interactive Inc. (Paramount Global (f/k/a ViacomCBS Inc.))
• Original Assignee: CBS Interactive Inc. (Paramount Global (f/k/a ViacomCBS Inc.))
• Inventors: Musgrove, Timothy A., Walsh, Robin Hiroko
• Granted Patent: US 7,363,214 B2
• Time in Patent Office: Days
• US Class Current: 715/256
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products

Abstract

A system and method for determining quality of written product reviews to distinguish the user reviews for further use or processing. In one embodiment, the opinion analyzer system includes a processor, a parser module, an prose analyzer module, a characteristic features database, and a language value rules database. In another embodiment, the method comprises the steps of parsing language of a portion of a user review, extracting characteristic feature from the user review, determining a quality parameter based on the extracted characteristic feature, determining a raw score based on the quality parameter, and determining quality of the user review based on the raw score.

Claims

1. A method for determining quality of user reviews to distinguish the user reviews for further use or processing, said method comprising the steps of:

parsing text of at least a portion of a user review;

extracting at least one characteristic feature from said user review that is indicative of quality of said user review;

determining at least one quality parameter for said user review based on said extracted characteristic feature;

determining a raw score for said at least one quality parameter; and

determining quality of said user review based on said raw score.

2. The method of claim 1, wherein said at least one characteristic feature is a text pattern.

3. The method of claim 2, wherein said text pattern is a wildcard based pattern having a wildcard therein.

4. The method of claim 1, wherein said step of determining at least one quality parameter based on said at least one characteristic feature includes determining a number of occurrences of said at least one characteristic feature.

5. The method of claim 1, wherein said step of determining said raw score for said at least one quality parameter includes the step of specifying an antecedent condition that is satisfied with a text pattern.

6. The method of claim 1, wherein said step of determining said raw score for said at least one quality parameter includes the step of counting each occurrence of said at least one characteristic feature in said user review.

7. The method of claim 1, wherein said step of determining said raw score for said at least one quality parameter includes the step of providing a predetermined ideal raw score.

8. The method of claim 7, further including the step of determining a functional score for said at least one quality parameter based on said raw score and said predetermined ideal raw score.

9. The method of claim 8, wherein said step of determining functional score includes the step of determining the difference between said raw score and said predetermined ideal raw score.

10. The method of claim 9, wherein said step of determining functional score further includes the step of assigning a functional value if said raw score corresponds to said ideal raw score.

11. The method of claim 8, wherein said step of determining functional score further includes the step of providing a predetermined high differential value, and a predetermined low differential value for said at least one characteristic feature.

12. The method of claim 11, wherein said step of determining functional score further includes the step of assigning zero value if the difference between said raw score and said predetermined ideal raw score is below said predetermined low differential value.

13. The method of claim 11, wherein said step of determining functional score further includes the step of assigning zero value if the difference between said raw score and said predetermined ideal raw score is above said predetermined high differential value.

14. The method of claim 11, wherein said step of determining functional score further includes the step of assigning an apportioned value if the difference between said raw score and said predetermined ideal raw score is above said predetermined low differential value and is below said predetermined high differential value, but is not zero.

15. The method of claim 14, wherein said apportioned value is apportioned proportionately based on said raw score relative to said predetermined idealized raw score and at least one of said predetermined high differential value and said predetermined low differential value.

16. The method of claim 8, further including the step of weighing said at least one quality parameter emphasize or de-emphasize said at least one quality parameter.

17. The method of claim 8, further including the step of determining a final quality score for said user review based on said functional score of each quality parameter of said user review.

18. The method of claim 17, further including the step of weighing at least one of said at least one quality parameter emphasize or de-emphasize said quality parameter to modify said final quality score.

19. The method of claim 1, further including the step of providing a fall-off factor that diminishes said raw score of said at least one quality parameter when said at least one characteristic feature is repeated.

20. The method of claim 1, wherein said at least one characteristic feature of said user review is a positively critical text pattern.

21. The method of claim 1, wherein said at least one characteristic feature of said user review is a negatively critical text pattern.

22. The method of claim 1, wherein said at least one characteristic feature of said user review is a nullifier text pattern nullifying effect of at least one of an positively critical text pattern and a negatively critical text pattern.

23. The method of claim 1, wherein said at least one characteristic feature of said user review is an enhancer text pattern enhancing effect of at least one of an positively critical text pattern and a negatively critical text pattern.

24. The method of claim 1, wherein said at least one characteristic feature of said user review is a comparative text pattern.

26. The method of claim 1, wherein said at least one characteristic feature of said user review is an experiential text pattern indicating real experience in usage of product reviewed by said user review.

27. The method of claim 1, wherein said at least one characteristic feature is a jargon text pattern associated with a product category.

28. The method of claim 1, further including the step of determining a balance parameter of said user review by calculating the difference between number of positively critical text patterns of said user review, and number of negatively critical text patterns of said user review.

29. The method of claim 1, further including the step of determining a depth parameter of said user review by determining a minimum of the number of positively critical text patterns and the number of negatively critical text patterns of said user review.

30. The method of claim 1, further including the step of determining a comparativeness parameter of said user review by determining number of comparative text patterns.

31. The method of claim 1, further including the step of determining an experiential parameter of said user review by determining number of experiential text patterns.

32. The method of claim 1, further including the step of determining a richness parameter of said user review by calculating a sum of positively critical text patterns, negatively critical text patterns, experiential text patterns, and comparative text patterns.

33. The method of claim 1, further including the step of determining a verbosity parameter of said user review by calculating ratio of length of said user review to sum of positively critical text patterns, negatively critical text patterns, comparative text patterns, and experiential text patterns.

34. The method of claim 1, further including the step of determining a conciseness parameter of said user review by calculating length of said user review in comparison to a predetermined length.

35. The method of claim 1, further including the step of determining a focus parameter of said user review by calculating length of text that discuss product of said user review in comparison to length of text that does not discuss product of said user review.

36. The method of claim 1, further including the step of determining a grammar parameter of said user review indicative of extent to which said user review avoids grammatical errors.

37. The method of claim 1, further including the step of determining a jargon parameter of said user review indicative of the number of technical jargons used in said user review.

38. The method of claim 17, further including the step of adjusting said final quality score of said user review based on information associated with a reviewer providing said user review.

39. The method of claim 38, wherein said information associated with said reviewer is at least one of an e-mail address associated with said user, responses to a questionnaire, recency of said user review, final quality scores of previous user reviews from said reviewer, and collaborative information.

40. The method of claim 1, wherein said user review is directed to at least one of a product and a service.

25. The method of claim 25, wherein said comparative text pattern is at least one of a brand name and a model of said another comparable product.

41. A system for determining quality of user reviews to distinguish the user reviews for further use or processing, said system comprising:

a processor;

a parser module adapted to parse text of at least a portion of a user review and extract a characteristic feature of said user review that is indicative of quality of said user review; and

a prose analyzer module adapted to determine at least one quality parameter based on said characteristic feature, determine a raw score for said at least one quality parameter, and to determine quality of said user review based on said raw score.

42. The system of claim 41, wherein said prose analyzer module determines said raw score for said at least one quality parameter by counting each occurrence of said characteristic feature in said user review.

43. The system of claim 41, further comprising a character features database having a plurality of character features stored therein.

44. The system of claim 43, wherein said plurality of character features are a plurality of text patterns.

45. The system of claim 44, wherein at least one of said plurality of text pattern is a wildcard based pattern having a wildcard therein.

46. The system of claim 44, further comprising a language value rules database having a plurality of rules specifying point assignment values associated with said plurality of text patterns.

47. The system of claim 41, wherein said prose analyzer module is further adapted to determine a functional score for said at least one quality parameter based on difference between said raw score and a predetermined ideal raw score.

48. The system of claim 47, wherein said prose analyzer module is further adapted to assign an apportioned value for said at least one quality parameter if the difference between said raw score and a predetermined ideal raw score is not within a predetermined differential value.

49. The system of claim 47, wherein said prose analyzer is further adapted to allow weighing of said at least one quality parameter to emphasize or de-emphasize said quality parameter to modify said final quality score.

50. The system of claim 47, wherein said prose analyzer module is further adapted to determine a final quality score for said user review based on said functional score for each quality parameter of said user review.

51. The system of claim 50, further comprising a submitter profiling module adapted to adjust said final quality score based on information associated with a reviewer providing said user review.

52. The system of claim 51, wherein said information associated with said reviewer is at least one of an e-mail address associated with said user, responses to a questionnaire, recency of said user review, final quality scores of previous user reviews from said reviewer, and collaborative information.

53. The system of claim 41, wherein said characteristic feature of said user review is at least one of a positively critical text pattern and a negatively critical text pattern.

54. The system of claim 41, wherein said characteristic feature of said user review is at least one of a nullifier text pattern and an enhancer text pattern.

55. The system of claim 41, wherein said characteristic feature of said user review is a comparative text pattern indicating a comparison between product reviewed by said user review and another comparable product.

56. The system of claim 41, wherein said characteristic feature of said user review is an experiential text pattern indicating real experience in usage of product reviewed by said user review.

57. The system of claim 41, wherein said characteristic feature of said user review is a jargon text pattern associated with a product category.

58. The system of claim 41, wherein said at least one quality parameter is a balance parameter which is the difference between number of positively critical text patterns of said user review, and number of negatively critical text patterns of said user review.

59. The system of claim 41, wherein said at least one quality parameter is a depth parameter which is a minimum of the number of positively critical text patterns and the number of negatively critical text patterns of said user review.

60. The system of claim 41, wherein said at least one quality parameter is a comparativeness parameter which is a number of comparative text patterns.

61. The system of claim 41, wherein said at least one quality parameter is an experiential parameter which is a number of experiential text patterns.

62. The system of claim 41, wherein said at least one quality parameter is a richness parameter which is a sum of positively critical text patterns, negatively critical text patterns, experiential text patterns, and comparative text patterns.

63. The system of claim 41, wherein said at least one quality parameter is a verbosity parameter which is a ratio of length of said user review to sum of positively critical text patterns, negatively critical text patterns, comparative text patterns, and experiential text patterns.

65. The system of claim 41, wherein said at least one quality parameter is a conciseness parameter which is the length of said user review in comparison to a predetermined length.

66. The system of claim 41, wherein said at least one quality parameter is a focus parameter which is the length of text that discuss product of said user review in comparison to length of text that does not discuss product of said user review.

67. The system of claim 41, wherein said at least one quality parameter is a grammar parameter indicative of extent to which said user review avoids grammatical errors.

68. The system of claim 41, wherein said at least one quality parameter is a jargon parameter indicative of the number of technical jargons used in the user review.

69. The system of claim 41, wherein said user review is directed to at least one of a product and a service.

70. A computer readable medium for determining quality of user reviews to distinguish the user reviews for further use or processing, said medium comprising:

instructions for parsing text of at least a portion of a user review;

instructions for extracting at least one characteristic feature from said user review that is indicative of quality of said user review;

instructions for determining at least one quality parameter for said user review based on said extracted characteristic feature;

instructions for determining a raw score for said at least one quality parameter; and

instructions for determining quality of said user review based on said raw score.

71. The medium of claim 70, wherein said instructions for determining at least one quality parameter includes instructions for determining a number of occurrences of said at least one characteristic feature.

72. The medium of claim 70, further including instructions for determining a functional score of said at least one quality parameter based on said raw score.

73. The medium of claim 72, further including instructions for determining a final quality score for said user review based on said functional score.

74. The medium of claim 73, further including instructions for weighing at least one of said at least one quality parameter emphasize or de-emphasize said quality parameter to modify said final quality score.

75. The medium of claim 70, further including instructions for diminishing said raw score of said at least one quality parameter when said at least one characteristic feature is repeated.

76. The medium of claim 73, further including instructions to allow adjustment of said final quality score of said user review based on information associated with a reviewer providing said user review.