BOOTSTRAPPING RECOMMENDER SYSTEM AND METHOD

• US 20120130931A1
• Filed: 11/18/2010
• Published: 05/24/2012
• Est. Priority Date: 11/18/2010
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Koren, Yehuda, Golbandi, Nadav, Lempel, Ronny
• Granted Patent: US 8,510,248 B2
• Time in Patent Office: Days
• US Class Current: 706/46
• CPC Class Codes: G06N 5/046 Forward inferencing; Production systems

Abstract

Bootstrapping a recommender system that makes item recommendations. The bootstrapping identifying questions for use in interviewing a user, e.g., a new user to the recommender system, to obtain user information, e.g., user profile information, for use in predicting item ratings for the user, the predicted item ratings for use in identifying item recommendations for the user. The bootstrapping using a cost function to minimize error in selection of the questions for the interview. The bootstrapping comprising a static bootstrapping and an adaptive bootstrapping.

Claims

1. A method comprising:

identifying, by at least one computing device, a set of interview items for an interview in which a user is requested to provide feedback, the identifying a set of interview items comprising, for each interview item identified;

determining, for each candidate item of a plurality of candidate items, a cost associated with including the candidate item in the set of interview items;

identifying the interview item using the cost associated with each of the candidate items of the plurality, the cost associated with the identified interview item being less than the cost associated with unselected candidate items of the plurality of candidate items;

removing the identified interview item from the plurality of candidate items;

presenting, by the at least one computing device, at least one of the interview items to a user and requesting feedback from the user about the item in at least one page of an interview user interface;

receiving, by the at least one computing device, the requested feedback from the user in response; and

identifying, by the at least one computing device, one or more items as recommendations using the feedback received from the user during the interview.

2. The method of claim 1, wherein the interview is adaptive, and wherein a second interview item of the set of interview items is presented in the adaptive interview subsequent to a first interview item of the set, the second interview item being presented based on the feedback received during the interview from the user.

3. The method of claim 2, wherein the user is asked to provide the feedback by selecting one of a plurality of options associated with the first interview item, wherein each interview item in the set of interview items corresponds to a node of a decision tree, the decision tree comprising a plurality of nodes, each node having a corresponding set of users and the cost associated with the corresponding interview item, and wherein the node is connected to another node via an edge representing one of the plurality of options.

4. The method of claim 3, further comprising:

in response to receiving the feedback for the first interview item, traversing the decision tree from a first node associated with the first interview item along an edge to reach a second node, the edge that is traversed corresponding to the one of the plurality of options selected by the user as the feedback for the first interview item, and the second node corresponding to the second interview item; and

presenting, in the interview user interface, the second interview item corresponding to the second node reached by traversing the decision tree along the edge from the first node to the second node.

5. The method of claim 3, identifying one or more items as recommendations using the feedback from the user to the interview items presented during the interview further comprising:

identifying the one or more items as recommendations using item ratings of the users associated with at least one node of the decision tree traversed during the interview.

6. The method of claim 1, wherein the interview is static, and wherein the at least one of the interview items is presented in the static interview independent of the feedback received from the user for another interview item presented in the static interview.

7. The method of claim 6, identifying one or more items as recommendations using the feedback from the user to the at least one of the interview items presented during the interview further comprising:

identifying the one or more items as recommendations using the feedback as ratings for the at least one of the interview items presented during the interview.

8. A system comprising:

at least one processor and memory, the memory storing instructions that when executed by the at least one processor cause the at least one processor to;

identify a set of interview items for an interview in which a user is requested to provide feedback, identifying a set of interview items comprising, for each interview item identified;

determining, for each candidate item of a plurality of candidate items, a cost associated with including the candidate item in the set of interview items;

identifying the interview item using the cost associated with each of the candidate items of the plurality, the cost associated with the identified interview item being less than the cost associated with unselected candidate items of the plurality of candidate items;

removing the identified interview item from the plurality of candidate items;

present at least one of the interview items to a user and request feedback from the user about the item in at least one page of an interview user interface;

receive the requested feedback from the user in response; and

identify one or more items as recommendations using the feedback received from the user during the interview.

9. The system of claim 8, wherein the interview is adaptive, and wherein a second interview item of the set of interview items is presented in the adaptive interview subsequent to a first interview item of the set, the second interview item being presented based on the feedback received during the interview from the user.

10. The system of claim 9, wherein the user is asked to provide the feedback by selecting one of a plurality of options associated with the first interview item, wherein each interview item in the set of interview items corresponds to a node of a decision tree, the decision tree comprising a plurality of nodes, each node having a corresponding set of users and the cost associated with the corresponding interview item, and wherein the node is connected to another node via an edge representing one of the plurality of options.

11. The system of claim 10, further comprising instructions that when executed by the at least one processor cause the at least one processor to:

in response to receiving the feedback for the first interview item, traverse the decision tree from a first node associated with the first interview item along an edge to reach a second node, the edge that is traversed corresponding to the one of the plurality of options selected by the user as the feedback for the first interview item, and the second node corresponding to the second interview item; and

present, in the interview user interface, the second interview item corresponding to the second node reached by traversing the decision tree along the edge from the first node to the second node.

12. The system of claim 10, the instructions to identify one or more items as recommendations using the feedback from the user to the at least one of the items presented during the interview further comprising instructions that when executed by the at least one processor cause the at least one processor to:

identify the one or more items as recommendations using item ratings of the users associated with at least one node of the decision tree traversed during the interview.

13. The system of claim 8, wherein the interview is static, and wherein the at least one of the interview items is presented in the static interview independent of the feedback received from the user for another interview item presented in the static interview.

14. The system of claim 13, the instructions to identify one or more items as recommendations using the feedback from the user to the items presented during the interview further comprising instructions that when executed by the at least one processor cause the at least one processor to:

identify the one or more items as recommendations using the feedback as ratings for the at least one of the interview items presented during the interview.

15. A computer readable storage media for tangibly storing thereon computer readable instructions that when executed, cause at least one processor to:

identify a set of interview items for an interview in which a user is requested to provide feedback, identifying a set of interview items comprising, for each interview item identified;

determining, for each candidate item of a plurality of candidate items, a cost associated with including the candidate item in the set of interview items;

identifying the interview item using the cost associated with each of the candidate items of the plurality, the cost associated with the identified interview item being less than the cost associated with unselected candidate items of the plurality of candidate items;

removing the identified interview item from the plurality of candidate items;

present at least one of the interview items to a user and request feedback from the user about the item in at least one page of an interview user interface;

receive the requested feedback from the user in response; and

identify one or more items as recommendations using the feedback received from the user during the interview.

16. The media of claim 15, wherein the interview is adaptive, and wherein a second interview item of the set of interview items is presented in the adaptive interview subsequent to a first interview item of the set, the second interview item being presented based on the feedback received during the interview from the user.

17. The media of claim 16, wherein the user is asked to provide the feedback by selecting one of a plurality of options associated with the first interview item, wherein each interview item in the set of interview items corresponds to a node of a decision tree, the decision tree comprising a plurality of nodes, each node having a corresponding set of users and the cost associated with the corresponding interview item, and wherein the node is connected to another node via an edge representing one of the plurality of options.

18. The media of claim 15, further comprising instructions that when executed, cause at least one processor to:

in response to receiving the feedback for the first interview item, traverse the decision tree from a first node associated with the first interview item along an edge to reach a second node, the edge that is traversed corresponding to the one of the plurality of options selected by the user as the feedback for the first interview item, and the second node corresponding to the second interview item; and

present, in the interview user interface, the second interview item corresponding to the second node reached by traversing the decision tree along the edge from the first node to the second node.

19. The media of claim 17, the instructions that when executed, cause at least one processor to identify one or more items as recommendations using the feedback from the user to the at least one of the items presented during the interview further comprising instructions that when executed, cause at least one processor to:

identify the one or more items as recommendations using item ratings of the users associated with at least one node of the decision tree traversed during the interview.

20. The media of claim 15, wherein the interview is static, and wherein the at least one of the interview items is presented in the static interview independent of the feedback received from the user for another interview item presented in the static interview.

21. The media of claim 20, instructions that when executed, cause at least one processor to identify one or more items as recommendations using the feedback from the user to the items presented during the interview further comprising instructions that when executed, cause at least one processor to:

identify the one or more items as recommendations using the feedback as ratings for the at least one of the interview items presented during the interview.