Recommendation system with multiple integrated recommenders

• US 8,751,507 B2
• Filed: 06/29/2007
• Issued: 06/10/2014
• Est. Priority Date: 06/29/2007
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Kim, Sung H., Lam, Shing Yan, Chakrabarti, Kushal, Ionkov, George M., Witt, Brett W.
• Primary Examiner(s): SYED, FARHAN M
• Application Number:
  US11/772,013
• Publication Number:
  US 20090006374A1
• Time in Patent Office: 2,538 Days
• Field of Search: 707/706, 707/722, 707/736, 707/758
• US Class Current: 707/748
• CPC Class Codes: G06F 16/335 Filtering based on additional data, e.g. user or group profiles filtering in web context G06F16/9535, G06F16/9536

Abstract

A recommendations system is provided in various embodiments for selecting items to recommend to a user. The system includes a recommendation engine with a plurality of recommenders, and each recommender identifies a different type of reason for recommending items. In one embodiment, each recommender retrieves item preference data and generates candidate recommendations responsive to a subset of that data. The recommenders also score the candidate recommendations. In certain embodiments, a normalization engine normalizes the scores of the candidate recommendations provided by each recommender. A candidate selector selects at least a portion of the candidate recommendations based on the normalized scores to provide as recommendations to the user. The candidate selector also outputs the recommendations with associated reasons for recommending the items.

Claims

1. A recommendations system for selecting items to recommend to a target user, the system comprising:

a computer system comprising computer hardware, the computer system programmed to implement;

a user interface that provides functionality for users of an electronic catalog of items to create and assign arbitrary textual tags to individual items represented in the electronic catalog after selecting said items, such that the textual tags categorize the items represented in the electronic catalog, the user interface comprising a tag entry field on item detail pages of the electronic catalog, the tag entry field enabling the users to create the tags via entry of text strings into the tag entry field;

a recommendation engine comprising a tag-based recommender configured to;

programmatically identify, from item preference data stored in computer storage, an electronic catalog item previously selected by a target user, without receiving a designation from the target user of the previously selected item,

subsequent to identification of the electronic catalog item, programmatically identify one or more tags of the textual tags associated with the previously selected catalog item in a data repository by at least identifying a threshold number of most popular tags assigned to the selected item, without any input from the target user, the one or more identified tags having been assigned to the electronic catalog item via the user interface, and

subsequent to said identification of the one or more tags, automatically search the electronic catalog using the one or more identified tags as one or more keywords to identify a corresponding set of related items; and

a candidate selector component configured to;

select at least a portion of the set of related items to provide as recommendations to the target user, and

output the recommendations with associated textual reasons for recommending the items.

2. The system of claim 1, wherein the electronic catalog item previously selected by the target user comprises an item purchased by the target user.

3. The system of claim 1, wherein the associated textual reason for recommending the items comprises text indicating that the items are similar to an item purchased by the user.

17. The system of claim 1, wherein the selected item comprises an item tagged by the target user.

18. The system of claim 1, wherein the computer system comprises multiple physical computers.

25. The system of claim 1, wherein the items of the electronic catalog are products that are available for purchase.

26. The system of claim 1, wherein the recommendation engine is further configured to calculate scores and rankings of the related items based on search result relevancy scores returned from the search of the electronic catalog.

27. The system of claim 1, wherein the candidate selector is further configured to include a reference to at least one of the one or more identified tags in the textual reasons for recommending the items.

28. The system of claim 26, wherein the candidate selector is further configured to select at least a portion of the set of related items based on the rankings of the related items.

4. A computer-implemented method of providing item recommendations to a target user, the method comprising:

by a computer system comprising computer hardware, the computer system configured to generate recommendations by at least;

by a computer system comprising computer hardware, the computer system configured to generate recommendations by at least;

providing a user interface that provides functionality for users of an electronic catalog of items to create and assign arbitrary textual tags to individual items represented in the electronic catalog after selecting said items, such that the textual tags categorize the items represented in the electronic catalog, the user interface comprising a tag entry field on item detail pages of the electronic catalog, the tag entry field enabling the users to create the tags via entry of text strings into the tag entry field;

programmatically identifying at least one tagged item that was selected by a target user, the at least one tagged item being represented in the electronic catalog, without receiving a designation from the target user of the at least one tagged item;

subsequent to and in response to said identifying, selecting one or more user-defined textual tags associated with the at least one tagged item by at least identifying a threshold number of most popular tags assigned to the at least one tagged item, without any input from the target user, each tag comprising a descriptor selectively associated with the at least one tagged item by one or more users;

subsequent to said selecting the one or more user-defined tags, automatically searching the electronic catalog using the one or more tags as one or more keywords to identify a corresponding set of related items;

selecting at least a portion of the set of related items to provide as the recommendations to the target user; and

outputting the recommendations with associated textual reasons for recommending the items.

5. The method of claim 4, wherein the at least one tagged item selected by the target user comprises an item purchased by the target user.

6. The method of claim 4, wherein the at least one tagged item selected by the target user comprises an item added to a wish list of the target user.

7. The method of claim 4, wherein the at least one tagged item selected by the target user comprises an item added to a shopping cart of the target user.

8. The method of claim 4, wherein the set of related items comprises lists of scored items, each list corresponding to one of the at least one tags.

9. The method of claim 8, further comprising merging the lists of scored items to generate a single list of ranked items.

10. The method of claim 8, further comprising adding scores of alike items appearing in different lists.

11. The method of claim 4, further comprising outputting the recommendations with an associated textual reason for recommending the items.

12. The method of claim 11, wherein the associated textual reason for recommending the items comprises text indicating that the items are similar to an item purchased by the user.

19. The method of claim 4, wherein the at least one tagged item selected by the target user comprises an item tagged by a user other than the target user.

29. The method of claim 4, further comprising limiting a scope of said searching to a category of items in the electronic catalog.

13. Non-transitory computer storage having instructions stored thereon which cause a computer system to perform a method of selecting items to recommend to a target user, the method comprising:

outputting a user interface that provides functionality for users of an electronic catalog of items to create and assign arbitrary textual tags to individual items represented in the electronic catalog after selecting said items, such that the textual tags categorize the items represented in the electronic catalog, the user interface comprising a tag entry field on item detail pages of the electronic catalog, the tag entry field enabling the users to create the tags via entry of text strings into the tag entry field;

programmatically identifying at least one tagged item in the electronic catalog, the at least one tagged item selected by a target user, without receiving a designation from the target user of the at least one tagged item;

subsequent to and in response to said identifying, selecting one or more user-defined textual tags associated with the at least one tagged item by at least identifying a threshold number of most popular tags assigned to the at least one tagged item, without any input from the target user, each tag comprising a descriptor selectively associated with the at least one tagged item by one or more users;

automatically searching the electronic catalog using the one or more tags as keywords to identify a corresponding set of related items;

selecting at least a portion of the set of related items to provide as recommendations to the target user; and

outputting the recommendations with associated textual reasons for recommending the items.

14. The non-transitory computer storage of claim 13, wherein the at least one tagged item selected by the target user is an item purchased by the target user.

20. The non-transitory computer storage of claim 13, in combination with a computer system that executes said instructions to perform said method.

21. The non-transitory computer storage of claim 13, wherein the at least one tagged item selected by the target user comprises an item tagged by the target user.

22. The non-transitory computer storage of claim 13, wherein the method further comprises outputting the recommendations with an associated textual reason for recommending the items.

15. An apparatus for selecting items to recommend to a target user, the apparatus comprising:

means for providing a user interface configured to enable a user of an electronic catalog of items to create textual tags for categorizing items represented in the electronic catalog and to associate particular tags with particular items in the electronic catalog, said means for providing comprising programmed computer hardware;

means for programmatically identifying at least one tagged item in the electronic catalog, the at least one tagged item selected by a target user, said means for identifying comprising programmed computer hardware, without receiving a designation from the target user of the at least one tagged item;

means for selecting one or more user-defined textual tags associated with the at least one tagged item subsequent to the identification of the at least one tagged item by at least identifying a threshold number of most popular tags assigned to the at least one tagged item, without any input from the target user, each tag comprising a descriptor selectively associated with the at least one tagged item by one or more users via the user interface, said means for selecting one or more user-defined textual tags comprising programmed computer hardware;

means for searching the electronic catalog, subsequent to the selection of the one or more user-defined textual tags, using the one or more tags as keywords to identify a corresponding set of related items, said means for searching comprising programmed computer hardware;

means for selecting at least a portion of the set of related items to provide as recommendations to the target user, said second means for selecting at least a portion of the set of related items comprising programmed computer hardware; and

means for outputting the recommendations with associated textual reasons for recommending the items.

16. The apparatus of claim 15, wherein the at least one tagged item selected by the target user comprises an item purchased by the target user.

23. The apparatus of claim 15, wherein the at least one tagged item selected by the target user comprises an item tagged by the target user.

24. The apparatus of claim 15, further comprising means for outputting the recommendations with an associated textual reason for recommending the items.