Relationship discovery engine

• US 8,005,724 B2
• Filed: 03/26/2003
• Issued: 08/23/2011
• Est. Priority Date: 05/03/2000
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Kindig, Bradley D, Dunning, Ted E
• Primary Examiner(s): Retta; Yehdega
• Application Number:
  US10/401,940
• Publication Number:
  US 20030229537A1
• Time in Patent Office: 3,072 Days
• Field of Search: None
• US Class Current: 705/26.7
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/40 of multimedia data, e.g. slideshows comprising image and additional audio data retrieval of still image data G06F16/50; retrieval of audio data G06F16/60; retrieval of video data G06F16/70
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0273 Determination of fees for advertising
• CPC Class Codes: G06Q 30/0277 Online advertisement
• CPC Class Codes: G06Q 30/0627 using item specifications
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G07F 17/16 for devices exhibiting advertisements, announcements, pictures or the like

Abstract

A system, method, and computer program product discover relationships among items and recommend items based on the discovered relationships. The recommendations provided by the present invention are based on user profiles that take into account actual preferences of users, without requiring users to complete questionnaires. An improved binomial log likelihood ratio analysis technique is applied, to reduce adverse effects of overstatement of coincidence and predominance of best sellers. The invention may be used, for example, to generate track lists for a personalized radio station.

Claims

1. A computer-implemented method comprising:

accepting, by at least one computer, input specifying at least one track;

selecting, by the at least one computer, a set of play logs from a corpus of play logs, each play log in the corpus corresponding to a user and being used to identify track occurrences from observed behavior of the user, each play log in the set is selected by taking into account a number of occurrences of the at least one track in the play log; and

selecting, by the at least one computer, a set of tracks for a playlist from a corpus of tracks comprising tracks found in the set of play logs selected, wherein selecting the set of tracks comprises analyzing track occurrences identified from observed behavior indicated in each play log in the selected set and track occurrences identified from observed behavior indicated in the corpus of play logs, each track in the selected set of tracks is determined to be overrepresented in the selected set of play logs based on a significance measurement for the track that takes into account multiple occurrences of the track in a play log.

2. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using a degree of co-occurrence of the track with the specified at least one track in the selected set of play logs.

3. The computer-implemented method of claim 1, wherein learned relationships exist among the tracks found in the selected set of play logs, and wherein selecting the set of tracks further comprises:

selecting the set of tracks using the learned relationships among tracks.

4. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a set of tracks responsive to parameters provided by one user.

5. The computer-implemented method of claim 1, wherein each track comprises a music track.

6. The computer-implemented method of claim 1, wherein each track comprises a video track.

7. The computer-implemented method of claim 1, further comprising transmitting, by the at least one computer, the set of tracks.

8. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises :

selecting a track to be a part of the set of tracks using information indicating a frequency with which the track has been played for one user.

9. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating a frequency with which the track has been played for a group of users.

10. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating an overall popularity of the track.

11. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating a popularity of the track within a demographic group.

12. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating how recently the track has been played for one user.

13. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating how recently the track has been played for a group of users.

14. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating that the track has not yet been played for one user.

15. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating that the track has not been played for a user within a predefined period of time.

16. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating user preferences.

17. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating user genre preferences.

18. The computer-implemented method of claim 1, wherein selecting the set of tracks further comprises:

selecting a track to be a part of the set of tracks using information indicating a detected user action with respect to the track.

19. The computer-implemented method of claim 18, wherein the detected user action comprises changing a volume level during playback of the track.

20. The computer-implemented method of claim 18, wherein the detected user action comprises skipping at least a portion of the track.

21. The computer-implemented method of claim 18, wherein the detected user action comprises repeating at least a portion of the track.

22. A computer-implemented method comprising:

accepting input specifying at least one artist;

selecting, by at least one computer, a set of play logs from a corpus of play logs, each play log in the corpus corresponding to a user and being used to identify artist occurrences from observed behavior of the user, each play log in the set is selected by taking into account a number of occurrences of the at least one artist found in the play log; and

selecting, by the at least one computer, a set of artists for a playlist from a corpus of artists comprising artists found in the set of play logs selected, wherein selecting the set of artists comprises analyzing artist occurrences identified from observed behavior indicated in each play log in the selected set and artist occurrences identified from observed behavior indicated in the corpus of play logs, each artist in the selected set of artists is determined to be overrepresented in the selected set of play logs based on a significance measurement for the artist that takes into account multiple occurrences of the artist in a play log.

23. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using a degree of co-occurrence of the artist with the specified at least one artist in the selected set of play logs.

24. The computer-implemented method of claim 22, wherein learned relationships exist among the artists found in the selected set of play logs, and wherein selecting the set of artists further comprises:

selecting the set of artists using the learned relationships among artists.

25. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting a set of artists responsive to parameters provided by one user.

26. The computer-implemented method of claim 22, wherein each artist comprises a musical artist.

27. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for one user.

28. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for a group of users.

29. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating an overall popularity.

30. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a popularity of the artist within a demographic group.

31. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for one user.

32. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for a group of users.

33. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not yet been played for one user.

34. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not been played for the user within a predefined period of time.

35. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user preferences.

36. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user genre preferences.

37. The computer-implemented method of claim 22, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a detected user action with respect to one or more tracks by the artist.

38. The computer-implemented method of claim 37, wherein the detected user action comprises changing a volume level during playback of the track by the artist.

39. The computer-implemented method of claim 37, wherein the detected user action comprises skipping at least a portion of the track by the artist.

40. The computer-implemented method of claim 37, wherein the detected user action comprises repeating at least a portion of the track by the artist.

41. A computer-implemented method comprising:

accepting, by at least one computer, input specifying at least one artist;

selecting, by the at least one computer, a set of play logs from a corpus of play logs, each play log in the corpus corresponding to a user and being used to identify artist occurrences from observed behavior of the user, each play log in the set is selected by taking into account a number of occurrences of the at least one artist in the play log; and

selecting, by the at least one computer, a set of artists for a playlist from a corpus of artists comprising artists found in the set of play logs selected, wherein selecting the set of artists comprises analyzing artist occurrences identified from observed behavior indicated in each play log in the selected set and artist occurrences identified from observed behavior indicated in the corpus of play logs, each artist in the selected set of artists is determined to be overrepresented in the set of play logs based on a significance measurement for the artist that takes into account multiple occurrences of the artist in the play log; and

selecting, by the at least one computer and for each of at least a subset of the artists, at least one track for the playlist.

42. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of set of artists using a degree of co-occurrence of the artist with the specified at least one artist in the selected set of play logs.

43. The computer-implemented method of claim 41, wherein learned relationships exist among the artists found in the selected set of play logs, and wherein selecting the set of artists further comprises:

selecting the set of artists using the learned relationships among artists.

44. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting a set of artists responsive to parameters provided by one user.

45. The computer-implemented method of claim 41, wherein each track comprises a music track.

46. The computer-implemented method of claim 41, wherein each track comprises a video track.

47. The computer-implemented method of claim 41, further comprising transmitting, by the at least one computer, the set of tracks.

48. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for one user.

49. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for a group of users.

50. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating an overall popularity.

51. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a popularity of the artist within a demographic group.

52. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for one user.

53. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for a group of users.

54. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not yet been played for one user.

55. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not been played for one user within a predefined period of time.

56. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user preferences.

57. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user genre preferences.

58. The computer-implemented method of claim 41, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a detected user action with respect to one or more tracks by the artist.

59. The computer-implemented method of claim 58, wherein the detected user action comprises changing a volume level during playback of a track by the artist.

60. The computer-implemented method of claim 58, wherein the detected user action comprises skipping at least a portion of a track by the artist.

61. The computer-implemented method of claim 58, wherein the detected user action comprises repeating at least a portion of a track by the artist.

62. A computer-implemented method comprising:

accepting, by at least one computer, input specifying at least one track;

determining, by the at least one computer, at least one artist for the specified at least one track;

selecting, by the at least one computer, a set of play logs from a corpus of play logs, each play log in the corpus corresponding to a user and being used to identify artist occurrences from observed behavior of the user, each play log in the set is selected by taking into account a number of occurrences of the at least one artist found in the play log; and

selecting, by the at least one computer, a set of artists for a playlist from a corpus of artists comprising artists found in the set of play logs selected, wherein selecting the set of artists comprises analyzing artist occurrences identified from observed behavior indicated in each play log in the selected set and artist occurrences identified from observed behavior indicated in the corpus of play logs, each artist in the selected set of artists is determined to be overrepresented in the selected set of play logs based on a significance measurement for the artist that takes into account multiple occurrences of the artist in the play log; and

selecting, by the at least one computer and for each of at least a subset of the set of selected artists, at least one track for the playlist.

63. The computer-implemented method of claim 62, wherein selecting the set of artists comprises:

selecting an artist to be a part of the set of artists using a degree of co-occurrence of the artist with the specified at least one artist in the selected set of play logs.

64. The computer-implemented method of claim 62, wherein learned relationships exist among the artists found in the selected set of play logs, and wherein selecting the set of artists further comprises:

selecting the set of artists using the learned relationships among artists.

65. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting the set of artists responsive to parameters provided by one user.

66. The computer-implemented method of claim 62, wherein each track comprises a music track.

67. The computer-implemented method of claim 62, wherein each track comprises a video track.

68. The computer-implemented method of claim 62, further comprising transmitting, by the at least one computer, the set of tracks.

69. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for one user.

70. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a frequency with which one or more tracks by the artist have been played for a group of users.

71. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating an overall popularity.

72. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a popularity of the artist within a demographic group.

73. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for one user.

74. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating how recently one or more tracks by the artist have been played for a group of users.

75. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not yet been played for one user.

76. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating that one or more tracks by the artist have not been played for one user within a predefined period of time.

77. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user preferences.

78. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating user genre preferences.

79. The computer-implemented method of claim 62, wherein selecting the set of artists further comprises:

selecting an artist to be a part of the set of artists using information indicating a detected user action with respect to a track by the artist.

80. The computer-implemented method of claim 79, wherein the detected user action comprises changing a volume level during playback of the track.

81. The computer-implemented method of claim 79, wherein the detected user action comprises skipping at least a portion of the track.

82. The computer-implemented method of claim 79, wherein the detected user action comprises repeating at least a portion of the track.

83. A computer-implemented method comprising:

accepting, by at least one computer, a set of preferences;

selecting, by the at least one computer, a set of play logs from a corpus of play logs, each play log in the corpus corresponding to a user and being used to identify track occurrences from observed behavior of the user, each play log in the set is selected by taking into account a number of occurrences of the at least one track in the play log; and

selecting, by the at least one computer, a set of tracks for a playlist from a corpus of tracks comprising tracks found in the set of play logs selected, wherein selecting the set of tracks comprises using the accepted set of preferences and analyzing track occurrences identified from observed behavior indicated in each play log in the selected set and track occurrences identified from observed behavior indicated in the corpus of play logs, each track in the selected set of tracks is determined to be overrepresented in the selected set of play logs based on a significance measurement for the track that takes into account multiple occurrences of the track in the play log.

84. The computer-implemented method of claim 83, wherein the user preferences comprise at least one selected from the group consisting of:

at least one genre;

at least one artist; and

at least one track.

85. The computer-implemented method of claim 83, further comprising:

determining, by the at least one computer, demographic information; and

wherein selecting the set of tracks further comprises selected the tracks using the determined demographic information.

86. The computer-implemented method of claim 83, wherein selecting the set of tracks further comprises selecting a track to be a part of the set of tracks using information indicating a frequency with which the track has been played for a group of users.

87. The computer-implemented method of claim 83, wherein selecting the set of tracks further comprises selecting a track to be a part of the set of tracks using information indicating an overall popularity of the track.

88. The computer-implemented method of claim 83, wherein selecting the set of tracks further comprises selecting a track to be a part of the set of tracks using information indicating a popularity within a demographic group.