Data mining system capable of generating pairwise comparisons of user-selectable items based on user event histories

• US 7,680,703 B1
• Filed: 06/05/2008
• Issued: 03/16/2010
• Est. Priority Date: 06/05/2008
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Smith, Brent R.
• Primary Examiner(s): Airapetian; Mila
• Application Number:
  US12/134,063
• Time in Patent Office: 649 Days
• Field of Search: 705/26, 705/27
• US Class Current: 705/26.8
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 30/06 Buying, selling or leasing transactions
• CPC Class Codes: G06Q 30/0633 Lists, e.g. purchase orders, compilation or processing

Abstract

Data mining systems and methods are disclosed for generating data that is helpful to users in selecting between items represented in an electronic data repository, such as an electronic catalog. One disclosed data mining method generates pairwise comparison data for particular pairs of items. The pairwise comparison data for a given item pair reveals the degree to which users who consider both items in the pair select one item over the other. The pairwise comparison data may be appropriately exposed to users of the electronic repository. For instance, in the context of an electronic catalog, an item detail page for item A may be supplemented with an indication that users who view both item A and item B select item B a specified percentage of the time. Another data mining method uses item viewing histories and item purchase histories of users in combination to identify pairs of items that are good candidates for being recommended in combination.

Claims

1. A system, comprising:

at least one data storage system that stores user activity data reflective of catalog item viewing events and catalog item acquisition events of each of a plurality of users of an electronic catalog;

a data mining system operative to analyze the user activity data reflective of the catalog item viewing events and catalog item acquisition events, and based thereon, to generate pairwise comparison data for each of a plurality of pairs of catalog items, said pairwise comparison data at least approximately indicating, for an item pair composed of a first catalog item and a second catalog item, how frequently users who view both the first catalog item and the second catalog item select the second catalog item to acquire over the first catalog item; and

a server system operative to incorporate the pairwise comparison data into catalog pages of the electronic catalog in association with catalog items to which such pairwise comparison data corresponds, the server system thereby enabling users to consider the pairwise comparison data in selecting catalog items to acquire.

2. The system of claim 1, wherein the data mining system is operative to generate the pairwise comparison data for said item pair at least partly by:

identifying a plurality of sessions in which both (1) the first catalog item and the second catalog item were both viewed, and (2) exactly one of the first and second catalog items was selected to acquire; and

determining how many of said plurality of sessions are sessions in which the second catalog item was selected to acquire.

3. The system of claim 2, wherein the data mining system is further operative to generate a percentage value reflective of how frequently the second catalog item was selected over the first catalog item in the plurality of sessions, and the server system is operative to incorporate said percentage value into an item detail page corresponding to the first catalog item.

4. The system of claim 2, wherein the data mining system is operative to give different non-zero amounts of weight to different ones of said sessions in generating the pairwise comparison data.

5. The system of claim 1, wherein the data mining system is operative to generate the pairwise comparison data for said item pair at least partly by:

identifying a set of users who both (1) viewed both the first catalog item and the second catalog, and (2) acquired exactly one of the first and second catalog items; and

determining how many of the users in said set acquired the second catalog item.

6. The system of claim 5, wherein the server system is operative to supplement an item detail page for the first item with an indication of what percentage of the users in said set acquired the second catalog item.

7. The system of claim 1, wherein the catalog item viewing events are item detail page viewing events.

8. The system of claim 1, wherein the catalog item acquisition events are item purchase events.

9. The system of claim 1, wherein the server system is operative to generate an item comparison page that includes pairwise comparison data for each of multiple pairs of catalog items.

10. The system of claim 9, wherein the item comparison page is an interactive page in which each of three or more catalog items is selectable to cause pairwise comparison data to be displayed which compares the selected catalog item on a pairwise basis to each of the other catalog items.

11. A system, comprising:

at least one data storage system that stores user activity data reflective of item consideration actions and item selection actions of each of a plurality of users, said item consideration actions and item selection actions corresponding to particular items represented in a browsable electronic data repository; and

a data mining system operative to analyze the user activity data reflective of the item consideration actions and the item selection actions of the plurality of users on an aggregated basis to generate pairwise comparison data for each of a plurality of item pairs, said pairwise comparison data at least approximately indicating, for an item pair composed of a first item and a second item, a degree to which users who consider both the first item and the second item select the second item over the first item, said data mining system further operative to store said pairwise comparison data in computer storage in association with the item pairs to which such pairwise comparison data corresponds.

12. The system of claim 11, further comprising a server system operative to expose the pairwise comparison data to users of the browsable electronic data repository in association with the items to which such pairwise comparison data corresponds.

13. The system of claim 11, wherein the data mining system is operative to generate the pairwise comparison data for said item pair at least partly by:

identifying a set of sessions in which both (1) the first item and the second item were both considered, and (2) exactly one of the first item and the second item was selected; and

determining how many of the sessions in said set are sessions in which the second item was selected.

14. The system of claim 11, wherein the data mining system is operative to generate the pairwise comparison data for said item pair at least partly by:

identifying a set of users that both (1) considered both the first item and the second item, and (2) selected the second item and not the first item; and

determining how many of the users in the set selected the second item.

15. The system of claim 11, wherein the items are catalog items represented in an electronic catalog.

16. The system of claim 15, wherein the item selection actions are catalog-item purchase actions.

17. The system of claim 11, wherein the items are users represented in a social networking system.

18. The system of claim 11, wherein the items are feeds to which users can subscribe, and the item selection actions are feed subscription actions.

19. A computer-implemented method of performing a pairwise comparison between a first item and a second item represented in an electronic data repository, the method comprising:

accessing session data stored in computer storage, said session data reflective of item consideration actions and item selection actions of each of a plurality of users, said item consideration actions and item selection actions corresponding to particular items represented in the electronic data repository;

based on the session data, identifying a set of sessions in which both (1) the first item and the second item were both considered, and (2) exactly one of the first item and the second item was selected, said set of sessions comprising a plurality of sessions of a plurality of users;

generating, by a computer system, pairwise comparison data based on the identified set of sessions, said pairwise comparison data indicating, at least approximately, a degree to which users who have considered both the first item and the second item have selected the second item over the first item; and

causing the pairwise comparison data to be presented to users of the electronic data repository in association with the first and second items.

20. The method of claim 19, wherein the electronic data repository is an electronic catalog, and the method comprises causing the pairwise comparison data to be presented on an item detail page of the electronic catalog.

21. The method of claim 19, wherein the items are travel destinations represented on a travel site, and the method comprises causing the pairwise comparison data to be presented on the travel site to assist users in selecting a travel destination.

22. The method of claim 19, wherein the pairwise comparison data indicates what percentage of users who considered both the first item and the second item selected the second item.

23. The method of claim 19, wherein generating the pairwise comparison data comprises giving different non-zero amounts of weight to different sessions in said set of sessions.

24. A computer-readable medium which stores executable instructions that embody the method of claim 19.

25. A computer system, including one or more physical computers coupled to computer storage, said one or more physical computers being programmed to perform the method of:

accessing session data stored in computer storage, said session data reflective of item consideration actions and item selection actions of each of a plurality of users, said item consideration actions and item selection actions corresponding to particular items represented in the electronic data repository;

based on the session data, identifying a set of sessions in which both (1) the first item and the second item were both considered, and (2) exactly one of the first item and the second item was selected, said set of sessions comprising a plurality of sessions of a plurality of users;

generating pairwise comparison data based on the identified set of sessions, said pairwise comparison data indicating, at least approximately, a degree to which users who have considered both the first item and the second item have selected the second item over the first item; and

causing the pairwise comparison data to be presented to users of the electronic data repository in association with the first and second items.