Automated content and collaboration-based system and methods for determining and providing content recommendations

• US 6,438,579 B1
• Filed: 07/14/2000
• Issued: 08/20/2002
• Est. Priority Date: 07/16/1999
• Status: Active Grant

• Current Assignee: Microsoft Corporation
• Original Assignee: AgentArts Incorporated (Microsoft Corporation)
• Inventors: Hosken, Benjamin E.
• Primary Examiner(s): Barot, Bharat
• Application Number:
  US09/616,474
• Time in Patent Office: 767 Days
• Field of Search: 709/200-203, 709/217-229, 705/7-10, 705/14, 705/26-27, 707/9-10, 707/1-6, 707/102-103, 235/375, 235/380, 235/462
• US Class Current: 709/203
• 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
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A content and collaborative filtering system for recommending entertainment oriented content items, such as music and video, and other media content items to a user based on similarity in profile between the user and other users and between the content indexed in the user'"'"'s profile and other content in the database. The system stores implicit and explicit ratings data for such content items provided by the users. Upon request of the user, the system accesses the user'"'"'s profile and corresponding content interests database. The system uses the relationships between the content items to determine a subset of the content items to be referred to the user. The system also correlates a similarity between the user'"'"'s ratings of the content items and other users'"'"' ratings. Based on the correlations, a subset of users is selected that is then used to provide recommendations to the user. The recommended items have a high probability of being subjectively appreciated by the user. The recommendations produced by the system will be represented to the user using a visual representation of the relationships between the content items allowing the user to explore the items related to the recommended items.

Claims

1. A content item referral system, executable by a computer, providing for the automated presentation of a set of recommended media-based content items in response to a query presented by a user, said content item referral system comprising:

a) a weighted relation subsystem operable to provide weighted relationships data representing relative similarities between characteristic attributes of a predetermined set of content items;

b) a referral sub-system, coupled to receive user profile data and said weighted relationship data, responsive to a user query, said referral system operative to perform a traversal of said user profile data and said weighted relationship data to provide an ordered list of content items relative to a predetermined content item; and

c) an action analysis sub-system coupled to said referral system to receive user action behaviors correlated to content items considered by said user, said action analysis sub-system providing said user profile data to said referral sub-system.

2. The content referral system of claim 1 wherein a user profile sub-system stores said user profile data in a form reflecting weighted relationships representing relative similarities between characteristic attributes of a user considered set of content items and wherein said referral sub-system is operable to combine said user profile data with said weighted relationship data.

3. The content referral system of claim 2 wherein said referral sub-system is operable as a graph traversal system over said combined user profile data and said weighted relationship data.

4. The content referral system of claim 3 wherein said user considered set of content items is a subset of said predetermined set of content items.

5. A method of providing media content recommendations through a computer server system connected to a network communications system, wherein said computer server system has access to a first database of media content items including media content and related information and a media content filter identifying and providing qualifying attribute relationship data for media content items within said first database, and wherein the media content recommendations are particularly tailored to the personalized interests of a user, said method comprising the steps of:

a) presenting media content items through a network-connected interface to a predetermined user for review and consideration of potential personal interest;

b) monitoring the consideration of said media content items implied through the user directed navigation among the presented media content items and user requests for related information;

c) collecting data from said step of monitoring to develop a user weighted data set reflective of said predetermined user'"'"'s relative consideration of said media content items; and

d) evaluating said user weighted data set in combination with said media content filter to identify a set of media content items accessible from said first database for presentation to said predetermined user consistent with said step of presenting.

6. The method of claim 5 wherein said step of collecting data further provides for the progressive refinement of said user weighted data set, whereby said step of evaluating progressively provides said set of media content items more closely tailored to the personalized interests of said predetermined user.

7. The method of claim 6 wherein said step of collecting data includes identifying the occurrence of predetermined actions taken by said predetermined user and determining selectively the durations of said predetermined actions relative to different media content items, and wherein said predetermined actions taken and the durations of said predetermined actions are incorporated into said user data set to reflect said predetermined user'"'"'s relative consideration of said media content items.

8. The method of claim 7 wherein said step of evaluating utilizes said user weighted data set to extrapolate through the relationships identified through said media content filter to identify said set of media content items.

9. The method of claim 8 wherein said step of evaluating operates as a graph traversal over the media content items related through said media content filter where the weighting data of said media content filter are modified to reflect said user weighted data set.

10. A content referral server system supporting, via a communications network, remote access, by a client system, to information relating, based on the similarity of characteristic attributes of specific instances of such content, different content items served by said content referral server system, said content referral server system comprising:

a) a content relations system that provides access to weighted content relationship information defining similarities between characteristic attributes of content referenceable by said content relations system; and

b) a profiling system that collects profiling information reflecting the navigational actions of a user of said client system in accessing said content referral server system, wherein said profiling system provides profile data combinable with said weighted content relationship information relative to content referenceable by said content relations system and selectable by said user of said client system.

11. The content referral server system of claim 10 wherein said profiling system provides profile data combinable with said weighted content relationship information relative to a characteristic attribute of the content selectable by said user of said client system.

12. The content referral server system of claim 11 wherein said profiling information reflects explicit indications of interest and implicit indications of interest in particular instances of content.

13. The content referral server system of claim 12 wherein said implicit indications of interest are derived from the selection and duration of sampling of particular instances of content.

14. A content item referral system, executable by a computer, providing for the automated presentation of a set of recommended media content items in response to a query presented by a user, said content item referral system comprising:

a) a first database storing weighted relationships data representing relative similarities between characteristic attributes of a predetermined set of content items;

b) a second database storing user profile data including weighted preferences with respect to a profile respective set of content items;

c) a referral generation system, coupleable to said first database and said second database to access said weighted relationship data and said user profile data, wherein said referral generation system is responsive to a user query to define a graph traversal of said weighted relationship data combined with said user profile data and qualified by a weighted rating and confidence level at predetermined graph traversal steps to provide an ordered list of content items responsive to sold user query and having a predetermined minimum weighted rating and confidence level.

15. The content item referral system of claim 14 wherein said content item referral system is interactively responsive to said user to receive said user query and user input representing user action behaviors and wherein said content item referral system further comprises a profile update system coupled to said second database to update user profile data corresponding to said user in response to user action behaviors correlated to content items considered by said user.

16. The content item referral system of claim 15 wherein said user action behaviors provide predetermined attributed data correlated to content items considered by said user, wherein an update weighted preference is determinable from said predetermined attributed data, and wherein said profile update system updates said second database with said updated weighted preference.

17. The content item referral system of claim 16 wherein said user action behaviors include navigational and content item selection actions.

18. The content item referral system of claim 17 wherein said attributed data defines an implicit level of interest by said user in a predetermined content item.

19. A content item referral system executable by a server computer system coupleable to a communications network and interactively responsive to user actions in connection with the review and selection of media content items through distributed client computer systems, including the sampling of media content items, to provide automated generation of recommended sets of media content items, said content item referral system comprising:

a) an expert database containing attributed expert weighting data defining a first node connected network describing a defined set of content items;

b) a user profile database containing a plurality of user profiles, wherein each said user profile includes attributed personalized weighting data for a respective subset of said defined set of content items, wherein said attributed personalized weighting data is derived from explicit and implicit user actions correlated to said plurality of user profiles, and wherein said attributed personalized weighing data defines a second node connected network;

c) a referral system, coupled to said expert database and said user profile database, responsive to a user query to provide a list of recommended media content items to evaluate a plurality of traversal paths through said first and second node connected networks qualified by a combination of said attributed expert weighting data and said attributed personalized weighting data having at least a determined minimum aggregate weighting, the terminal nodes of said traversal paths representing said list of recommended media content items.

20. The content item referral system of claim 19 wherein said referral system is responsive to group behavioral data reflecting explicit and implicit user actions correlated to content items within said set of content items.

21. The content item referral system of claim 20 further comprising a weighting filter coupled between said expert database and said referral system, said weighting filter providing for the normalization of said attributed expert weighting data with respect to said attributed personalized weighting data.

22. The content item referral system of claim 21 wherein the normalization provided by said weighting filter is derived from said group behavioral data.