Detection of behavior-based associations between search strings and items

• US 8,543,584 B2
• Filed: 02/06/2012
• Issued: 09/24/2013
• Est. Priority Date: 02/13/2006
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Vadon, Eric R., Whitman, Ronald M., Kohavi, Ron, Jayaraman, Gautam K., Redman, Benjamin W. S.
• Primary Examiner(s): VU, BAI DUC
• Application Number:
  US13/367,134
• Publication Number:
  US 20120143840A1
• Time in Patent Office: 596 Days
• Field of Search: None
• US Class Current: 707/751
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A system and method are disclosed for automatically detecting associations between particular sets of search criteria, such as particular search strings, and particular items. Actions of users of an interactive system, such as a web site, are monitored over time to generate event histories reflective of searches, item selection actions, and possibly other types of user actions. An analysis component collectively analyzes the event histories to automatically identify and quantify associations between specific search strings (or other types of search criteria) and specific items.

Claims

1. A computer-implemented method of determining a degree of association between a search string and an item, the computer-implemented method comprising:

as implemented by one or more computing devices configured with specific executable instructions,

retrieving a user event history sequence that identifies a number of user actions that occurred during user interaction with an electronic repository, wherein the user event history sequence comprises a search event in which the user submitted a request to search items in the electronic repository;

identifying a search string submitted by the user in association with the search event;

identifying a subsequent item selection event in the user event history sequence, wherein the subsequent item selection event corresponds to a selection by the user of an item in the electronic repository, said subsequent item selection event occurring after the search event;

determining a number of intervening events in the user event history sequence that each occur after the search event and before the subsequent item selection event; and

calculating a relationship value representing a probability that a user who submits the search string will subsequently select the item, wherein the relationship value is calculated based at least in part on the determined number of intervening events that occur after the search event and before the subsequent item selection event in the user event history sequence.

2. The computer-implemented method of claim 1, wherein the relationship value is calculated based at least in part on a plurality of user event history sequences associated with the search string.

3. The computer-implemented method of claim 2, wherein the plurality of user event history sequences comprises a user event history sequence corresponding to a different user.

4. The computer-implemented method of claim 1, wherein the relationship value is calculated based at least in part on a type of intervening event that occurred after the search event and before the subsequent item selection event.

5. The computer-implemented method of claim 1, wherein each intervening event serves to reduce an amount of weight given to the item selection event for purposes of calculating the relationship value.

6. The computer-implemented method of claim 1, wherein at least one of the intervening events includes a second search event, and wherein calculating the relationship value comprises taking into consideration a degree to which respective search strings of the search event and the second search event are related.

7. The computer-implemented method of claim 1, wherein the relationship value is calculated based at least in part on an amount of time that passed between the search event and the subsequent item selection event.

8. The computer-implemented method of claim 1, wherein the item is a product represented in an electronic catalog.

9. The computer-implemented method of claim 1, wherein the item is a category of products represented in an electronic catalog.

10. A system for determining a degree of association between a search string and an item, the system comprising:

a data store configured to store user event history sequences that identify a number of user actions that occurred during user interaction with an electronic repository, wherein at least one of the user event history sequences comprises a search event in which a user submitted a request to search items in the electronic repository; and

a computing device in communication with the data store, the computing device configured to;

retrieve the at least one user event history sequence from the data store, wherein the at least one user event history sequence includes a search string submitted by the user in association with the search event and a subsequent item selection event, wherein the subsequent item selection event corresponds to a selection by the user of an item in the electronic repository, said subsequent item selection event occurring after the search event;

determine a number of intervening events in the at least one user event history sequence that each occur after the search event and before the subsequent item selection event; and

calculate an event weight that specifies an amount of weight to give to the item selection event for purposes of quantifying an association between the search string submitted by the user and the item selected in the subsequent item selection event, wherein the event weight is calculated based at least in part on the determined number of intervening events that occur after the search event and before the subsequent item selection event in the at least one user event history sequence; and

based at least in part on the event weight, determine a relationship value representing a probability that a user who submits the search string will subsequently select the item.

11. The system of claim 10, wherein the relationship value is further determined based at least in part on event weights calculated for other item selection events associated with the item and the search string.

12. The system of claim 10, wherein, when the intervening events include an intervening search event, the event weight is calculated based at least in part on a degree of association between the search string and a second search string submitted in the intervening search event.

13. The system of claim 10, wherein the event weight is calculated based at least in part on an amount of time between the search event and the item selection event.

14. The system of claim 10, wherein the event weight is calculated based at least in part on an event type of at least one of the intervening events.

15. The system of claim 10, wherein the items are products represented in an electronic catalog.

16. The system of claim 10, wherein the items are web sites accessed via a search engine.

17. A non-transitory computer-readable medium comprising computer-executable components for determining a degree of association between a search string and an item, the computer-executable components comprising:

an identification component that identifies a user event history sequence that identifies a number of user actions that occurred during user interaction with an electronic repository, wherein the user event history sequence comprises;

a search event in which the user submitted a request to search items in the electronic repository, the search event associated with a search string submitted by the user in association with the search event;

a subsequent item selection event corresponding to a selection by the user of an item in the electronic repository, said subsequent item selection event occurring after the search event; and

a number of intervening events in the user event history sequence that each occur after the search event and before the subsequent item selection event; and

a calculation component that calculates a relationship value representing a probability that a user who submits the search string will subsequently select the item, wherein the relationship value is calculated based at least in part on the number of intervening events that occur after the search event and before the subsequent item selection event in the user event history sequence.

18. The non-transitory computer-readable medium of claim 17, wherein the relationship value is calculated based at least in part on a plurality of user event history sequences associated with the search string.

19. The non-transitory computer-readable medium of claim 17, further comprising a search engine component that uses the relationship value to determine whether to augment a query result set with the item.

20. The non-transitory computer-readable medium of claim 19, wherein the search engine component inhibits the item from being added to the query result set if a popularity level of the item exceeds a threshold.