COLLABORATIVE FILTERING-BASED RECOMMENDATIONS

• US 20080010258A1
• Filed: 06/19/2007
• Published: 01/10/2008
• Est. Priority Date: 06/22/2006
• Status: Active Grant

• Current Assignee: Infosys Limited
• Original Assignee: Infosys Technologies Limited (Infosys Limited)
• Inventors: Sureka, Ashish
• Granted Patent: US 8,468,155 B2
• Time in Patent Office: Days
• US Class Current: 707/003
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

Various techniques can be used to implement a collaborative filtering-based recommendation engine. For example, different similarity measures can be used for different users. Different similarity measures can be used for a particular user across time. A superior similarity measure can be found for a user. User-defined similarity measures can be supported.

Claims

1. A computer-implemented method comprising:

choosing item recommendations for different users via different user similarity measures when applying collaborative filtering for the different users; and

outputting the item recommendations.

2. The method of claim 1 further comprising:

selecting a similarity measure for a particular user via a stored mapping between users and user similarity measures.

3. The method of claim 1 further comprising:

choosing item recommendations for a same user via different user similarity measures when applying collaborative filtering for the same user at different times while choosing item recommendations for an other user via a same user similarity measure.

4. The method of claim 1 wherein the choosing comprises:

choosing a first recommendation for a first user via collaborative filtering using a first user similarity measure; and

choosing a second recommendation for a second user via a second user similarity measure different from the first.

5. The method of claim 1 further comprising:

selecting a user similarity measure out of a plurality of user similarity measures for use when choosing item recommendations for a user based on performance of the user similarity measure.

6. The method of claim 5 wherein the selecting comprises:

testing a plurality of similarity measures;

selecting a best performing similarity measure; and

storing an association associating the similarity measure with the user.

7. The method of claim 1 further comprising:

receiving from a user an indication regarding a preferred similarity measure; and

employing the preferred similarity measure when choosing recommendations for the user.

8. The method of claim 1 further comprising:

receiving from a user a user-defined similarity measure; and

employing the user-defined similarity measure when choosing item recommendations for the user.

9. The method of claim 8 wherein the user-defined similarity measure comprises a plurality of user-defined weight values provided.

10. The method of claim 1 wherein the choosing is performed in a single recommender system.

11. The method of claim 1 wherein the choosing is performed in a single recommendation engine.

12. One or more computer-readable media having encoded thereon computer-executable instructions causing a computer to perform the method of claim 1.

13. A computer-implemented method comprising:

choosing item recommendations for different users via different user similarity measures when applying collaborative filtering for the different users;

choosing item recommendations for a first user via different user similarity measures when applying collaborative filtering for the first user at different times while choosing item recommendations for a second user via a same user similarity measure;

selecting a user similarity measure out of a plurality of user similarity measures for use when choosing item recommendations for a third user based on performance of the user similarity measure;

receiving from a fourth user a user-defined similarity measure; and

employing the user-defined similarity measure when generating item recommendations for the fourth user; and

outputting the item recommendations.

14. One or more computer-readable media having encoded thereon a data structure comprising:

an indication of a user; and

an indication of a similarity function associated with the user and designated for application when determining a neighborhood of similar users in a recommender system configured to provide item recommendations for the user.

15. The one or more computer-readable media of claim 14 wherein the data structure comprises indications of a plurality of users and indications of similarity functions designated to be applied when determining neighborhoods of similar users for respective of the plurality of users.

16. The one or more computer-readable media of claim 14 wherein the similarity function is a user-defined function defined by the user.

17. One or more computer-readable media having encoded thereon a recommender system comprising:

a plurality of user ratings for a plurality of items;

a user similarity measure mapping configured to map different users to different similarity measures; and

a collaborative filtering engine configured to apply the different similarity measures for the different users against the plurality of user ratings for the plurality of items and generate recommendations therefrom.

18. The one or more computer-readable media of claim 17 wherein the recommender system is configured to switch among similarity measures for a single user over time.

19. The one or more computer-readable media of claim 17 wherein the recommender system further comprises:

a similarity measure evaluator configured to selected one out of a plurality of similarity measures for a user based on similarity measure performance for the user.

20. The one or more computer-readable media of claim 17 wherein the different similarity measures comprise at least one user-defined similarity measure.

21. An apparatus comprising:

means for storing a a plurality of user ratings for a plurality of items;

means for storing a user similarity measure mapping configured to map different users to different similarity measures; and

means for collaborative filtering configured to apply the different similarity measures for the different users against the plurality of user ratings for the plurality of items and generate recommendations therefrom.

22. A computer-implemented method comprising:

testing a plurality of user similarity measures for a particular user, wherein the testing determines respective predictive accuracies of the user similarity measures for the particular user;

selecting a best performing user similarity measure out of the user similarity measures based on predictive accuracy of the user similarity measure; and

designating the best performing user similarity measure as to be used when generating recommendations for the particular user with collaborative filtering.