Computer processes for identifying related items and generating personalized item recommendations

• US 8,024,222 B2
• Filed: 08/20/2007
• Issued: 09/20/2011
• Est. Priority Date: 09/18/1998
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon.com, Inc.
• Inventors: Jacobi, Jennifer A., Benson, Eric A., Linden, Gregory D.
• Primary Examiner(s): CHAMPAGNE, DONALD
• Application Number:
  US11/841,971
• Publication Number:
  US 20080033746A1
• Time in Patent Office: 1,492 Days
• Field of Search: 705/14.53, 705/14.67, 705/14.73, 705/26.7
• US Class Current: 705/14.53
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0253 During e-commerce, i.e. online transactions
• CPC Class Codes: G06Q 30/0254 based on statistics
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0271 Personalized advertisement
• CPC Class Codes: G06Q 30/0277 Online advertisement
• CPC Class Codes: G06Q 30/0282 Rating or review of business operators or products
• 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
• CPC Class Codes: G06Q 30/0633 Lists, e.g. purchase orders, compilation or processing
• CPC Class Codes: G06Q 30/0641 Shopping interfaces
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

A computer-implemented service analyzes purchase histories and/or other types of behavioral data of users on an aggregated basis to detect and quantify associations between particular items represented in an electronic catalog. The detected associations are stored in a mapping structure that maps items to related items, and is used to recommend items to users of the electronic catalog.

Claims

1. A computerized method of selecting items to recommend to users, the method comprising:

recording item viewing activities of each of a plurality of users of a system that provides network-based user access to a data repository of items, said item viewing activities reflecting user interests in particular items;

generating, by a computer system, an item-to-item mapping structure that maps each of a plurality of the items to a respective set of similar items, wherein generating the item-to-item mapping structure comprises analyzing the recorded item viewing activities of the plurality of users to assess degrees to which particular items are similar to each other, and generating, for each of a plurality of pairs of items, a respective data value that represents the degree to which the two items of that pair are similar to each other, said data values reflecting the analysis of the recorded item viewing activities;

receiving input data specifying a plurality of items selected by a user;

for each of the plurality of items selected by the user, looking up, from the item-to-item mapping structure, a respective set of similar items;

generating, by a computer system, a ranked set of similar items, said ranked set of similar items comprising the plurality of sets of similar items looked up from the item-to-item mapping structure; and

selecting, from the ranked set of similar items, a subset of items to recommend to the user.

2. The method of claim 1, wherein the input data specifying a plurality of items selected by the user is based on recorded item viewing activities of said user.

3. The method of claim 1, wherein the data repository is an electronic catalog of products, and wherein monitoring item viewing activities comprises recording events in which users view particular products in the electronic catalog.

4. The method of claim 1, further comprising storing a selected subset of the data values in the item-to-item mapping structure in association with the items to which such data values correspond.

5. The method of claim 1, wherein generating the ranked set of similar items comprises merging the looked-up sets of similar items while combining data values associated with like items.

6. The method of claim 1, wherein generating the ranked set of similar items comprises using the associated data values as a factor for ranking the similar items relative to each other.

7. The method of claim 1, wherein generating the ranked set of similar items comprises giving more weight to a first of said sets of similar items than to a second of said sets of similar items.

8. The method of claim 1, wherein generating the ranked set of similar items comprises, for a set of similar items corresponding to an item explicitly rated by the user, weighting the set of similar items by an amount that depends on the user'"'"'s explicit rating of said explicitly rated item.

9. The method of claim 1, wherein the item-to-item mapping structure is additionally based on explicit user ratings of particular items.

10. The method of claim 1, wherein the items are videos.

11. The method of claim 1, wherein the method comprises generating the item-to-item mapping structure in an off-line processing mode prior to receiving said input data.

12. A recommendation system, comprising:

a computer system comprising one or more physical servers, said computer system configured to at least;

record item viewing activities of each of a plurality of users of a system that provides network-based access to an item data repository, said item viewing activities reflecting user interests in particular items;

generate an item-to-item mapping that maps individual items to respective sets of similar items, wherein generating the item-to-item mapping comprises analyzing the recorded item viewing activities of the plurality of users to assess degrees to which particular items are similar to each other, and generating, for each of a plurality of pairs of items, a respective data value that represents the degree to which the two items of that pair are similar to each other, said data values reflecting the analysis of the recorded item viewing activities;

receive input data specifying a plurality of items selected by a user;

for each of the plurality of items selected by the user, look up, from the item-to-item mapping, a respective set of similar items;

generate a ranked set of similar items, said ranked set of similar items comprising the plurality of sets of similar items looked up from the item-to-item mapping; and

select, from the ranked set of similar items, a subset of items to recommend to the user.

13. The recommendation system of claim 12, wherein the input data specifying a plurality of items selected by the user is based on recorded item viewing activities of said user.

14. The recommendation system of claim 12, wherein the computer system is configured to record the item viewing activities by recording events in which users view products in an electronic catalog.

15. The recommendation system of claim 12, wherein the computer system is further configured to include a selected subset of the data values in the item-to-item mapping in association with the items to which such data values correspond.

16. The recommendation system of claim 12, wherein the computer system is configured to generate the ranked set of similar items by a process that comprises merging the looked-up sets of similar items while combining data values associated with like items.

17. The recommendation system of claim 12, wherein the computer system is configured to generate the ranked set of similar items by a process that comprises using the associated data values as a factor for ranking the similar items relative to each other.

18. The recommendation system of claim 12, wherein the computer system is configured to generate the item-to-item mapping based additionally on explicit user ratings of particular items.

19. The recommendation system of claim 12, wherein the items are videos.

20. The recommendation system of claim 12, wherein the computer system comprises an off-line process that is configured to generate the item-to-item mapping in an off-line processing mode.

21. The recommendation system of claim 20, wherein the computer system further comprises a recommendation process that is configured to use the item-to-item mapping to generate personalized item recommendations substantially in real time in response to requests from online users.

22. Physical computer storage that stores executable code that directs a computing system to implement a method that comprises:

recording video viewing activities of each of a plurality of users of a system that provides network-based access to a data repository of videos, said video viewing activities reflecting user interests in particular videos;

generating a video-to-video mapping that maps individual videos to respective sets of related videos, wherein generating the video-to-video mapping comprises analyzing the recorded video viewing activities of the plurality of users to assess degrees to which particular videos are related to each other, and generating, for each of a plurality of pairs of videos, a respective data value that represents the degree to which the two videos of that pair are related to each other, said data values reflecting the analysis of the recorded video viewing activities;

receiving input data specifying a plurality of videos selected by a user;

for each of the plurality of videos selected by the user, looking up, from the video-to-video mapping, a respective set of related videos;

generating a ranked set of related videos, said ranked set of related videos comprising the plurality of sets of related videos looked up from the video-to-video mapping; and

selecting, from the ranked set of related videos, a subset of videos to recommend to the user.

23. The physical computer storage of claim 22, wherein the executable code directs the computing system to identify the plurality of videos based on video viewing activities of the user.

24. The physical computer storage of claim 22, in combination with said computing system, wherein the computing system is programmed with said executable code to implement said method.

25. A recommendation system, comprising:

a computer system comprising one or more computers, said computer system configured to at least;

store data representing item viewing activities of each of a plurality of users, said item viewing activities reflective of item interests of the users;

generate, based at least partly on the item viewing activities of the plurality of users, data values that represent degrees to which specific items are related to each other, each data value corresponding to a respective pair of items, and representing a degree to which the two items of that pair are related to each other;

use the data values to select, for each of a plurality of items, a corresponding set of related items;

generate a mapping structure that maps each of said plurality of items to the respective set of related items; and

use the mapping structure, in combination with corresponding ones of said data values, to generate personalized item recommendations for each of a plurality of users.

26. The recommendation system of claim 25, wherein the computer system is configured to identify a first plurality of items viewed by a target user, and to use the mapping structure and associated data values to map the first plurality of items to a second plurality of items to recommend to the target user.