System and process for improving product recommendations for use in providing personalized advertisements to retail customers

• US 8,924,265 B2
• Filed: 01/14/2013
• Issued: 12/30/2014
• Est. Priority Date: 05/06/2008
• Status: Active Grant

• Current Assignee: RichRelevance Incorporated
• Original Assignee: RichRelevance Incorporated
• Inventors: Selinger, David Lee, Kohn, Tyler David, DeCourcey, Michael, Ahuja, Sundeep, Osial, James, Sunwoo, Albert
• Primary Examiner(s): JOHNSON, GREGORY L
• Application Number:
  US13/741,251
• Publication Number:
  US 20130198007A1
• Time in Patent Office: 715 Days
• Field of Search: None
• US Class Current: 705/26.7
• CPC Class Codes: G06Q 30/00 Commerce
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]

Abstract

A system and process for improving product recommendations for a first user includes receiving a request for one or more product recommendations for a first user, each product recommendation being associated with any one of a plurality of retailers, receiving a plurality of recommendation sets from one or more automated product recommendation systems, wherein the plurality of recommendation sets are generated using different selection models and using ensemble learning to select one or more most relevant product recommendation sets from the plurality of product recommendation sets.

Claims

1. A computer-implemented method comprising:

receiving, by a configured computing system, a request for one or more product recommendations to provide to an indicated user based at least in part on an information page for display to the indicated user, the information page having information about one or more products available from a retailer that are associated with one or more product categories and with one or more brands;

obtaining a plurality of product recommendation sets that are generated using multiple distinct selection algorithms;

using ensemble learning to select one of the multiple selection algorithms based at least in part on prior user behavior, and selecting one or more product recommendation sets from the plurality based on the selected one selection algorithm, wherein the selected one or more product recommendation sets are a subset of the plurality and each includes indications of multiple products that are recommended by one of the multiple selection algorithms, and wherein the using of the ensemble learning and the selecting of the one or more product recommendation sets are performed by the configured computing system in response to the received request and are based at least in part on the one or more product categories and on the one or more brands associated with the one or more available products; and

providing, by the configured computing system, an indication of one or more of the multiple products included in the selected one or more product recommendation sets, to enable information about at least one of the indicated one or more products to be provided for display to the indicated user.

2. The method of claim 1 wherein the prior user behavior occurred subsequent to previous product recommendations generated using one or more of the multiple selection algorithms, and wherein the selecting of the one or more product recommendation sets includes selecting one product recommendation set of the plurality based on the selected one product recommendation set being generated using the selected one selection algorithm.

3. The method of claim 1 wherein the multiple selection algorithms are part of multiple selection models, wherein the method further comprises comparing, by the configured computing system, prior performance of the multiple selection models by analyzing the prior user behavior, and wherein the selecting of the one selection algorithm is further based at least in part on the comparing of the prior performance.

4. The method of claim 1 wherein the selecting of the one or more product recommendation sets is further based at least in part on predictions of future user behavior from a Bayesian network.

5. The method of claim 4 wherein the Bayesian network is a natï

Bayesian network that utilizes information concerning the retailer, the information page, and the indicated user.

6. The method of claim 1 wherein the providing of the indication of the one or more products included in the selected one or more product recommendation sets includes electronically communicating the selected product recommendation sets.

7. The method of claim 1 wherein the information page is a Web page from a retail website of the retailer, and wherein the providing of the indication of the one or more products included in the selected one or more product recommendation sets further includes initiating display of advertising for the indicated one or more products on a copy of the Web page displayed to the indicated user.

8. The method of claim 1 wherein the information page is part of a retail website of the retailer, and wherein the providing of the indication of the one or more products included in the selected one or more product recommendation sets further includes initiating display of advertising for the indicated one or more products in one or more locations that are not part of the retail website.

9. The method of claim 1 further comprising generating a selection model using one of the multiple selection algorithms based at least in part on collected product information that includes at least one of the one or more product categories, and using the generated selection model to generate one of the selected one or more product recommendation sets.

10. The method of claim 1 further comprising generating a selection model using one of the multiple selection algorithms based at least in part on collected product information that includes at least one of the one or more brands, and using the generated selection model to generate one of the selected one or more product recommendation sets.

11. The method of claim 1 wherein the selected one or more product recommendation sets include a product recommendation set that is generated by a selection model using one of the multiple selection algorithms, the selection model being based at least in part on collected product information that includes at least one of the one or more product categories.

12. The method of claim 1 wherein the selected one or more product recommendation sets include a product recommendation set that is generated by a selection model using one of the multiple selection algorithms, the selection model being based at least in part on collected product information that includes at least one of the one or more brands.

13. The method of claim 1 wherein the configured computing system is part of a system for providing targeted content that is distinct from the retailer, wherein the received request includes information about one or more contextual attributes of the information page, and wherein the providing of the indication of the one or more products further includes providing advertising that is determined based at least in part on the contextual attributes.

14. The method of claim 13 wherein the information page is a Web page from a retail website of the retailer, and wherein the one or more contextual attributes include a Uniform Resource Locator (URL) of the Web page, a page/site hierarchy for the Web page, or information about a learned context based on historical user interactions with the Web page.

15. The method of claim 1 wherein the configured computing system is part of a system for providing targeted content that is distinct from the retailer, wherein the received request includes information about one or more profile attributes of the indicated user, and wherein the providing of the indication of the one or more products further includes providing advertising that is determined based at least in part on the profile attributes.

16. The method of claim 15 wherein the information page is a Web page from a retail website of the retailer, and wherein the one or more profile attributes include information about a history of pages viewed by the indicated user, information about a history of purchases by the indicated user, or information about tracked activity of the indicated user across multiple distinct websites.

17. The method of claim 1 wherein the obtaining of the plurality of product recommendation sets, the using of the ensemble learning to select one of the multiple selection algorithms, and the selecting of the one or more product recommendation sets are based at least in part on the one or more product categories and on the one or more brands and on a current context of the indicated user.

18. A non-transitory computer-readable medium having stored computer-executable instructions that when executed configure a computer system to perform a method, the method comprising:

receiving, by the configured computer system, a request for one or more recommendations to provide to an indicated user based at least in part on information for display to the indicated user, the information indicating one or more products that are associated with one or more product categories or with one or more brands;

obtaining information about a plurality of recommendation sets that are available to be generated using multiple selection algorithms, and selecting a recommendation set from the plurality that is generated by one of the multiple selection algorithms based at least in part on prior user behavior regarding the one selection algorithm, wherein the obtaining and selecting is performed by the configured computer system and is based at least in part on the one or more product categories or on the one or more brands; and

providing, by the configured computer system, an indication of at least one product included in the selected recommendation set, to enable information about the at least one product to be provided to the indicated user.

19. The non-transitory computer-readable medium of claim 18 wherein the selecting of the recommendation set includes selecting the one selection algorithm using ensemble learning based at least in part on the prior user behavior, and selecting the recommendation set based on the selected recommendation set being generated using the selected one selection algorithm.

20. The non-transitory computer-readable medium of claim 18 wherein the information indicating the one or more products is a Web page from a retail website of a retailer, and wherein the providing of the indication of the at least one product further includes initiating display of advertising for the indicated at least one product on a copy of the Web page displayed to the indicated user.

21. The non-transitory computer-readable medium of claim 19 wherein the one or more products indicated in the information are associated with an indicated product category, and wherein the selected one selection algorithm is part of a selection model based at least in part on collected product information that includes the indicated product category.

22. The non-transitory computer-readable medium of claim 19 wherein the one or more products indicated in the information are associated with an indicated brand, and wherein the selected one selection algorithm is part of a selection model based at least in part on collected product information that includes the indicated brand.

23. The non-transitory computer-readable medium of claim 18 wherein the received request includes information about one or more contextual attributes of a Web page from a retailer that includes the information indicating the one or more products, wherein the configured computer system is part of a system for providing targeted content that is distinct from the retailer, and wherein the providing of the indication of the at least one product further includes providing advertising that is determined based at least in part on the contextual attributes.

24. The non-transitory computer-readable medium of claim 18 wherein the received request includes information about one or more profile attributes of the indicated user, wherein the configured computer system is part of a system for providing targeted content that is distinct from a source that provides the information indicating the one or more products, and wherein the providing of the indication of the at least one product further includes providing advertising that is determined based at least in part on the profile attributes.

25. A system comprising:

a memory; and

a controller associated with the memory, wherein the controller is configured to;

receive a request for one or more recommendations to provide to an indicated user based at least in part on information for display to the indicated user, the information indicating one or more products that are associated with one or more product categories or with one or more brands;

obtain information about a plurality of recommendation sets that are available to be generated using multiple selection algorithms, and selecting a recommendation set from the plurality that is generated by one of the multiple selection algorithms, wherein the selecting is performed in response to the received request and is based at least in part on the one or more product categories or on the one or more brands; and

provide an indication of at least one product included in the selected recommendation set, to enable information about the at least one product to be provided to the indicated user.

26. The system of claim 25 wherein the selecting of the recommendation set includes selecting the one selection algorithm using ensemble learning based at least in part on prior user behavior, and selecting the recommendation set based on the selected recommendation set being generated using the selected one selection algorithm.

27. The system of claim 25 wherein the information indicating the one or more products is a Web page from a retail website of a retailer, and wherein the providing of the indication of the at least one product further includes initiating display of advertising for the indicated at least one product on a copy of the Web page displayed to the indicated user.

28. The system of claim 26 wherein the one or more products indicated in the information are associated with an indicated product category, and wherein the selected one selection algorithm is part of a selection model based at least in part on collected product information that includes the indicated product category.

29. The system of claim 26 wherein the one or more products indicated in the information are associated with an indicated brand, and wherein the selected one selection algorithm is part of a selection model based at least in part on collected product information that includes the indicated brand.

30. The system of claim 25 wherein the received request includes information about one or more profile attributes of the indicated user and about one or more contextual attributes of a Web page from a retailer that includes the information indicating the one or more products, and wherein the providing of the indication of the at least one product further includes providing advertising that is determined based at least in part on the profile attributes and on the contextual attributes.