Content recommendation device with an arrangement engine

• US 7,716,220 B2
• Filed: 06/04/2004
• Issued: 05/11/2010
• Est. Priority Date: 06/04/2003
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: RealNetworks, Inc.
• Inventors: Zich, Otto, Schramm, Balthasar, Petro, Oliver, Michelitsch, Stefan, Riegler, Andreas, Spechtler, Andreas, Nemetz, Thomas, Burgstaller, Rainer
• Primary Examiner(s): Breene; John E
• Assistant Examiner(s): Gofman; Alex
• Application Number:
  US10/861,133
• Publication Number:
  US 20050021420A1
• Time in Patent Office: 2,167 Days
• Field of Search: 707/104.1
• US Class Current: 707/736
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

The invention relates to a content recommendation device comprising a selection engine (2) that selects content items from a content item pool (4) and generates a selection list (9) comprising the selected content items, and/or means to receive a selection list The content recommendation device also comprises an arrangement engine (3) that generates a content recommendation list (12) by re-arranging at least a part of content items of the selection list (9) so that preferably a dramaturgic sequence of content items is generated.

Claims

1. A content recommendation device, comprising:

a processor;

a computer readable storage medium;

a selection engine configured to select songs and/or videos from a pool and generate a selection list including the selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values, or means for receiving a selection list including selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values; and

an arrangement engine configured to generate a content recommendation list by;

generating a plurality of clusters of songs and/or videos, each cluster being comprised of songs and/or videos sharing a different respective value for a first song and/or video attribute;

generating, within each cluster, a plurality of sub-clusters of songs and/or videos, each sub-cluster being comprised of songs and/or videos sharing a different respective value for a second song and/or video attribute;

selecting, based on arrangement rules, a sequence of songs and/or videos from both at least one cluster and at least one sub-cluster to include in the content recommendation list, wherein the sequence defines an order in which the respective songs and/or videos are played back to a user, and wherein the arrangement rules define;

an order in which to access the at least one cluster and the at least one sub-cluster during selection of the sequence, wherein the order defines accessing a first cluster or sub-cluster before accessing a second cluster or sub-cluster;

how many songs and/or videos to select from the first cluster or sub-cluster before accessing the second cluster or sub-cluster; and

how many songs and/or videos to select from the second cluster or sub-cluster, wherein the rules specify that two or more songs and/or videos are selected from at least one of the first cluster or sub-cluster and the second cluster or sub-cluster; and

providing arranged content delivery by outputting a stream of songs and/or videos arranged in the selected sequence, wherein the stream is configured to be sequentially reproduced in the arranged order.

2. The content recommendation device according to claim 1, wherein said arrangement engine generates said content recommendation list based on said arrangement rules such that said sequence comprises a dramaturgic sequence of songs and/or videos.

3. The content recommendation device according to claim 1, wherein said arrangement engine generates said content recommendation list based on song and/or video attributes that define every song and/or video in a standardized way.

4. The content recommendation device according to claim 2, wherein said arrangement rules combine individual song and/or video attributes.

5. The content recommendation device according to claim 1, wherein said arrangement engine generates said content recommendation list based on business rules that generate a commercially influenced sequence of songs and/or videos.

6. The content recommendation device according to claim 1, wherein said selection engine generates said selection list based on song and/or video attributes that define every song and/or video in a standardized way.

7. The content recommendation device according to claim 1, wherein said selection engine generates said selection list based on business rules that generate a commercially influenced sequence of songs and/or videos.

8. The content recommendation device according to claim 1, wherein said arrangement engine generates said content recommendation list based on a user profile.

9. The content recommendation device according to claim 8, wherein said user feedback is used to generate an iteratively refined recommendation list on basis of a reduced content item pool that includes the songs and/or videos of the selection list, or a new selection list.

10. The content recommendation device according to claim 1, wherein said selection engine selects the songs and/or videos from a pool according to a user profile.

11. The content recommendation device according to claim 1, further comprising:

a user profiling engine configured to generate a user profile based on a user feedback to at least one previously output recommendation list.

12. The content recommendation device according to claim 11, wherein said user feedback includes explicit and implicit feedback.

13. The content recommendation device according to claim 12, wherein said user feedback is a function of a representation of the explicit feedback weighted with a representation of the implicit feedback.

14. The content recommendation device according to claim 12, wherein said explicit feedback includes an indication whether a user likes or dislikes a particular song and/or video.

15. The content recommendation device according to claim 12, wherein said implicit feedback is based on the level of reaction accompanying the explicit feedback.

16. The content recommendation device according to claim 15, wherein said implicit feedback is based on the intensity the user is operating an input terminal of the content recommendation device.

17. The content recommendation device according to claim 15, wherein said implicit feedback includes a pressure put on an explicit feedback control.

18. The content recommendation device according to claim 15, wherein said implicit feedback includes a reaction time of a user to operate an explicit feedback control.

19. The content recommendation device according to claim 1, wherein the number of clusters is equal to the number of different possible values for the first attribute.

20. The content recommendation device according to claim 1, wherein the number of sub-clusters is equal to the number of different possible values for the second attribute.

21. The content recommendation device according to claim 1, wherein each of the plurality of sub-clusters comprises a first order sub-cluster, the method further comprising:

generating, within each first order sub-cluster, a plurality of second order sub-clusters of songs and/or videos, each second order sub-cluster being comprised of songs and/or videos sharing a different respective value for a third song and/or video attribute.

22. The content recommendation device according to claim 21, further comprising:

selecting, based on one or more arrangement rules, a number of songs and/or videos to include in the content recommendation list from at least one second order sub-cluster in addition to the songs and/or videos selected from the at least one cluster and the at least one first order sub-cluster.

23. The content recommendation device according to claim 1, further comprising:

generating, within each N^th order sub-cluster, a plurality of (N+1)^th order sub-clusters of songs and/or videos, each (N+1)^th order sub-cluster being comprised of songs and/or videos sharing a different respective value for an (N+2)^th song and/or video attribute.

24. The content recommendation device according to claim 23, wherein the generating of (N+1)^th order sub-clusters repeats until (O−

1)^th order sub-clusters are generated, where O equals the number of individual song and/or video attributes.

25. The content recommendation device according to claim 1, wherein at least one arrangement rule relies on information from a user profile to determine whether to select a song and/or video from a particular Nth order sub-cluster.

26. The content recommendation device according to claim 1, wherein at least one arrangement rule relies on an indication of whether the user liked or disliked a previously presented song and/or video to determine whether to select a song and/or video from a particular Nth order sub-cluster.

27. A method of content recommendation, comprising:

selecting songs and/or videos from a content item pool and generating a selection list of the selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values, or receiving a selection list including selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values;

generating a plurality of clusters of songs and/or videos, each cluster being comprised of songs and/or videos sharing a different respective value for a first song and/or video attribute;

generating, within each cluster, a plurality of sub-clusters of songs and/or videos, each sub-cluster being comprised of songs and/or videos sharing a different respective value for a second song and/or video attribute;

selecting, based on arrangement rules, a sequence of songs and/or videos from both at least one cluster and at least one sub-cluster to include in a content recommendation list, wherein the sequence defines an order in which the respective songs and/or videos are played back to a user, and wherein the arrangement rules define;

an order in which to access the at least one cluster and the at least one sub-cluster during selection of the sequence, wherein the order defines accessing a first cluster or sub-cluster before accessing a second cluster or sub-cluster;

how many songs and/or videos to select from the first cluster or sub-cluster before accessing the second cluster or sub-cluster; and

how many songs and/or videos to select from the second cluster or sub-cluster, wherein the rules specify that two or more songs and/or videos are selected from at least one of the first cluster or sub-cluster and the second cluster or sub-cluster; and

providing arranged content delivery by outputting a stream of songs and/or videos arranged in the selected sequence, wherein the stream is configured to be sequentially reproduced in the arranged order.

28. The method according to claim 27, wherein generating said content recommendation list is based on said arrangement rules such that said sequence comprises a dramaturgic sequence of songs and/or videos.

29. The method according to claim 28, further comprising:

combining individual song and/or video attributes with said arrangement rules.

30. The method according to claim 27, wherein generating said content recommendation list is based on song and/or video attributes that define every song and/or video in a standardized way.

31. The method according to claim 27, wherein generating said content recommendation list is based on business rules that generate a commercially influenced sequence of songs and/or videos.

32. The method according to claim 27, wherein generating said selection list is based on song and/or video attributes that define every song and/or video in a standardized way.

33. The method according to claim 27, wherein generating said selection list is based on business rules that generate a commercially influenced sequence of songs and/or videos.

34. The method according to claim 27, wherein generating said content recommendation list is based on a user profile.

35. The method according to claim 27, wherein selecting the songs and/or videos from a content item pool is according to a user profile.

36. The method according to claim 27, wherein generating a user profile is based on a user feedback to at least one previously output recommendation list.

37. The method according to claim 36, wherein said user feedback includes explicit and implicit feedback.

38. The method according to claim 37, wherein user feedback is a function of a representation of the explicit feedback weighted with a representation of the implicit feedback.

39. The method according to claim 37, wherein said explicit feedback includes an indication whether a user likes or dislikes a particular song and/or video.

40. The method according to claim 37, wherein said implicit feedback is based on the level of reaction accompanying the explicit feedback.

41. The method according to claim 40, wherein said implicit feedback is based on the intensity the user is operating an input terminal of the content recommendation device.

42. The method according to claim 40, wherein said implicit feedback includes a pressure put on an explicit feedback control.

43. The method according to claim 40, wherein said implicit feedback includes a reaction time of a user to operate an explicit feedback control.

44. The method according to claim 36, wherein said user feedback is used to generate an iteratively refined recommendation list on basis of a reduced content item pool that includes the songs and/or videos of the selection list, or a new selection list.

45. A computer readable storage medium including computer program instructions stored thereon that cause a computer to execute a method of content recommendation comprising:

selecting songs and/or videos from a content item pool and generating a selection list of the selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values, or receiving a selection list including selected songs and/or videos, each song and/or video being associated with one or more song and/or video attributes having a number of possible values;

generating a plurality of clusters of songs and/or videos, each cluster being comprised of songs and/or videos sharing a different respective value for a first song and/or video attribute;

generating, within each cluster, a plurality of sub-clusters of songs and/or videos, each sub-cluster being comprised of songs and/or videos sharing a different respective value for a second song and/or video attribute;

selecting, based on arrangement rules, a sequence of songs and/or videos from both at least one cluster and at least one sub-cluster to include in the content recommendation list, wherein the sequence defines an order in which the respective songs and/or videos are played back to a user, and wherein the arrangement rules define;

an order in which to access the at least one cluster and the at least one sub-cluster during selection of the sequence, wherein the order defines accessing a first cluster or sub-cluster before accessing a second cluster or sub-cluster; and

how many songs and/or videos to select from the first cluster or sub-cluster before accessing the second cluster or sub-cluster; and

how many songs and/or videos to select from the second cluster or sub-cluster, wherein the rules specify that two or more songs and/or videos are selected from at least one of the first cluster or sub-cluster and the second cluster or sub-cluster; and

providing arranged content delivery by outputting a stream of songs and/or videos arranged in the selected sequence, wherein the stream is configured to be sequentially reproduced in the arranged order.