Customer keyword-based item recommendations

• US 9,390,168 B1
• Filed: 09/28/2010
• Issued: 07/12/2016
• Est. Priority Date: 09/28/2010
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Brozovich, Stephen Michael, Days, Adam Lloyd, Dykstra, Aaron James, Cao, Tri M., Vora, Jinesh U.
• Primary Examiner(s): Chen, Susan
• Application Number:
  US12/891,931
• Time in Patent Office: 2,114 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/686 using information manually generated, e.g. tags, keywords, comments, title or artist information, time, location or usage information, user ratings
• CPC Class Codes: G06F 16/903 Querying for retrieval from the web G06F16/953
• CPC Class Codes: G06F 16/9035 Filtering based on additional data, e.g. user or group profiles
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]
• CPC Class Codes: G06Q 30/0603 Catalogue ordering
• CPC Class Codes: G06Q 30/0631 Item recommendations

Abstract

Disclosed are various embodiments for generating recommendations based at least in part on keywords associated with users. In some embodiments, among others, a system includes at least one computing device and a recommendation generator executable in the at least one computing device. The recommendation generator comprises logic that generates a plurality of pools of keywords based at least in part on a plurality of behavioral histories. Each pool corresponds to a behavioral history of a user across a plurality of domains. The recommendation generator also comprises logic that clusters at least a number of the keywords in a cluster across at least two pools including the same keyword and logic that recommends an item based at least in part on the cluster of keywords.

Claims

1. A non-transitory computer-readable medium embodying a program executable in a computing device, the program comprising:

code that gathers content based at least in part on view data associated with a user on a first domain;

code that identifies a plurality of keywords based at least in part on the content;

code that associates the keywords and the user to form keyword-user associations;

code that assigns a score to each of the keywords corresponding to the keyword-user associations;

code that ranks the keywords based at least in part on the corresponding score;

code that selects a number of the keywords based at least in part on a ranking;

code that retrieves a plurality of other keyword-user associations based at least in part on the selected keywords, the other keyword-user associations corresponding to other users and a second domain; and

code that recommends an item available from the first domain based at least in part on the other keyword-user associations.

2. The computer-readable medium of claim 1, wherein each of the other keyword-user associations is associated with a score, the program further comprising code that ranks the other keyword-user associations based at least in part on the score.

3. The computer-readable medium of claim 1, wherein the recommendation is further based at least in part on keyword-keyword associations.

4. A system, comprising:

at least one computing device; and

a recommendation generator executable in the at least one computing device, the recommendation generator comprising;

logic that generates a plurality of pools of keywords based at least in part on a plurality of behavioral histories, each pool corresponding to a behavioral history of a user across a plurality of domains;

logic that clusters at least a number of the keywords in a cluster across at least two pools including the same keyword; and

logic that recommends an item based at least in part on the cluster of keywords.

5. The system of claim 4, wherein the recommendation generator further comprises logic that ranks the keywords in each pool.

6. The system of claim 4, further comprising logic that assigns a score to each keyword in each pool based at least in part on the corresponding behavioral history.

7. The system of claim 4, wherein the recommendation generator further comprises logic that ranks the keywords in each pool based at least in part on the score.

8. The system of claim 4, wherein the recommendation is based at least in part on the cluster.

9. The system of claim 4, further comprising logic that identifies at least one keyword that is included in the at least two pools, wherein the at least one keyword is also highly ranked in each of the at least two pools.

10. The system of claim 4, further comprising logic that selects a number of highly ranked keywords from each of the at least two pools, the number of highly ranked keywords from each of the at least two pools being included in the cluster.

11. A method, comprising the steps of:

identifying keywords based at least in part on a behavioral history corresponding to a user;

associating the keywords with the user to form keyword-user associations; and

recommending an item to the user based at least in part on the keyword-user associations.

12. The method of claim 11, wherein the behavioral history includes a first purchase history corresponding to a first domain and a second purchase history corresponding to a second domain.

13. The method of claim 11, wherein the behavioral history includes a view history gathered from a plurality of domains.

14. The method of claim 11, wherein the behavioral history includes content gathered from a social networking forum.

15. The method of claim 14, wherein the content gathered from the social networking forum includes content related to the user.

16. The method of claim 14, wherein the content gathered from the social networking forum includes content related to contacts associated with the user.

17. The method of claim 11, further comprising the step of assigning a score to each of the keyword-user associations.

18. The method of claim 11, further comprising the step of ranking the keyword-user associations based at least in part on the respective scores, the recommendation being further based at least in part on the ranked keyword-user associations.

19. The method of claim 11, wherein the score of the keyword-user associations are based at least in part on at least one of the following factors corresponding to the respective keyword:

a number of occurrences in the behavioral history, a phrase length, a part of speech, a purchase, a view, a return, and a selection.

20. The method of claim 11, wherein the score is based at least in part on at least one of a time decay and an inverse general population keyword popularity.

21. The method of claim 11, wherein recommending the item is also based at least in part on a keyword-keyword association.

22. The method of claim 11, wherein recommending the item is also based at least in part on a keyword-item association.

23. The method of claim 11, further comprising:

encoding the recommendation in a network page; and

sending the network page to a client.