Personalization using multiple personalized selection algorithms

• US 20080133298A1
• Filed: 12/14/2007
• Published: 06/05/2008
• Est. Priority Date: 08/13/2003
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Clyde D. Mcqueen, Michael Irvin Hyman, Richard Scott Brinkerhoff
• Inventors: McQueen, Clyde D., Hyman, Michael Irvin, Brinkerhoff, Richard Scott
• Granted Patent: US 7,765,130 B2
• Time in Patent Office: Days
• US Class Current: 705/7
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0245 Surveys
• CPC Class Codes: G06Q 30/0254 based on statistics
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G06Q 30/0623 Item investigation
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: G06Q 30/0641 Shopping interfaces
• CPC Class Codes: G06Q 50/10 Services

Abstract

A system provides an interface for users to author and post pieces of content, referred to as “blurbs,” for viewing by other users. Blurbs may also be obtained from external sources, such as from blogs hosted by various web sites. A personalized blurb selection component selects blurbs to present to users based on histories of catalog items selected by such users, and/or based on various other criteria. User feedback provided on specific blurbs may be taken into consideration by the personalized blurb selection algorithms.

Claims

1-33. -33. (canceled)

34. A system for selecting items to present to users, the system comprising:

a plurality of item selectors, each of which uses a different respective personalized selection algorithm to select items to nominate for presentation to a target user, each item selector configured to output item nominations together with a relevance score for each nominated item; and

an arbiter that receives the item nominations and associated relevance scores from the item selectors, normalizes the relevance scores for the nominated items, and uses at least the normalized relevance scores to select particular nominated items to present to the target user.

35. The system of claim 34, wherein the items include blurbs authored by users.

36. The system of claim 34, wherein the system is implemented as part of an electronic catalog system that hosts a browsable electronic catalog of items, and at least some of the item selectors use catalog-item selection histories of particular users to nominate items to present to such users.

37. The system of claim 34, wherein at least one of the item selectors uses different information about the target user than another one of said item selectors to select items to nominate.

38. The system of claim 34, wherein at least some of the item selectors use item-specific feedback from users to select the items to nominate.

39. The system of claim 34, wherein at least some of the item selectors use an item selection history of the target user to select particular items to nominate.

40. The system of claim 34, wherein the arbiter selects particular nominated items to present to the target user at least in part by comparing the normalized relevance scores.

41. The system of claim 34, wherein the arbiter, in determining whether to select a nominated item to the target user, takes into consideration whether more than one of the item selectors nominated said nominated item for the target user.

42. The system of claim 34, wherein the arbiter selects particular nominated items to present to the target user so as to ensure that the items presented to the target user are not dominated by the nominations of a single item selector.

43. The system of claim 34, wherein the arbiter selects particular nominated items to present to the target user so as to ensure that the target user is presented with items nominated by each of the item selectors.

44. The system of claim 34, wherein the arbiter uses feedback from target users on particular items to assess, on a user-specific basis, how well particular item selectors perform for particular target users.

45. The system of claim 34, wherein the arbiter is responsive to the target user consistently responding favorably to items nominated by a particular item selector by favoring nominations of the particular item selector for the target user.

46. The system of claim 34, further comprising a personal log component that creates a personalized log for each user containing items select by the arbiter for the respective user.

47. A computer-implemented method of selecting items to present to users, the method comprising:

using a plurality of personalized selection algorithms to select items to nominate for presentation to a target user, each personalized selection algorithm outputting one or more item nominations for the target user together with an associated relevance score for each nominated item, said personalized selection algorithms using different respective selection criteria to select particular items to nominate; and

receiving the item nominations and associated relevance scores from the personalized selection algorithms, normalizing the relevance scores, and using at least the normalized relevance scores to select particular items to present to the target user.

48. The method of claim 47, wherein the items include blurbs authored by users.

49. The method of claim 47, wherein the method is implemented in association with an electronic catalog system that hosts a browsable electronic catalog of items, and at least some of the personalized selection algorithms use information about catalog-item selections of particular users to nominate items to present to the users.

50. The method of claim 47, wherein the method comprises using personalized item selection algorithms that use different respective information about the target user to select items to nominate.

51. The method of claim 47, wherein at least some of the personalized selection algorithms use item-specific feedback from users to select the items to nominate.

52. The method of claim 47, wherein using at least the normalized relevance scores to select particular items to present comprises comparing the normalized relevance scores.

53. The method of claim 47, wherein the method comprises selecting from the nominated items so as to ensure that the items selected to present to the target user are not dominated by nominations of a particular personalized selection algorithm.

54. The method of claim 47, wherein the method comprises selecting from the nominated items so as to ensure that the target user is presented with items nominated by each of the item selectors.

55. The method of claim 47, wherein the method comprises using feedback from the target user on particular items selected to present to the target user to assess how well particular personalized selection algorithms perform for the target user.

56. The method of claim 47, wherein the method comprises responding to consistent favorable feedback from the target user on items nominated by a particular personalized selection algorithm by increasing a frequency with which items nominated by the particular personalized selection algorithm are presented to the target user.

57. The method of claim 47, further comprising inserting the items selected to present to the target user into a personal log of the target user.

58. Computer storage which stores executable code that embodies the method of claim 47.