Strategy for providing query results based on analysis of user intent

• US 7,860,886 B2
• Filed: 09/29/2006
• Issued: 12/28/2010
• Est. Priority Date: 09/29/2006
• Status: Active Grant

• Current Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Original Assignee: A9.com Incorporated (Amazon.com, Inc.)
• Inventors: Loftesness, David E.
• Primary Examiner(s): Truong; Cam Y
• Assistant Examiner(s): Chau; Dung K
• Application Number:
  US11/537,004
• Publication Number:
  US 20080082518A1
• Time in Patent Office: 1,551 Days
• Field of Search: 707/3, 707/5, 707/100, 705/5, 705/10, 705/14
• US Class Current: 707/771
• CPC Class Codes: G06F 16/90335 Query processing
• CPC Class Codes: G06Q 10/02 Reservations, e.g. for tickets, services or events
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A strategy is described for responding to a user'"'"'s query based on a consideration of the user'"'"'s intent in entering the query. The user'"'"'s intent, in turn, is determined by examining prior query-related behavior of a population of users.

Claims

1. A computerized method for providing query results based at least in part on user intent, comprising:

receiving, at a computing device, a query from each of a plurality of users, each query of queries received from the plurality of users being a same query;

generating and providing query results to each user of the plurality of users at least in part in response to the receiving of the query;

identifying actions taken by said each user of the plurality of users upon receiving the query results to determine query-related behavior of the plurality of users after submitting the query by identifying patterns in the actions that evince a predominant intent of the users in entering the queries;

associating a result-generating function with the query based at least in part on the determined query-related behavior of the plurality of users after submitting the query;

receiving the query from an individual user, the query received from the individual user being the same query previously submitted by said each user of the plurality of users;

identifying, at least in part in response to the receiving of the query from the individual user and using classification mapping information, the result-generating function associated with the query, wherein the classification mapping information maps queries into different result-generating functions based on an analysis of prior query-related behavior of a plurality of users; and

generating query results for the individual user using the identified result-generating function that is based at least in part on the determined query-related behavior of the plurality of users that have submitted the query.

2. The computerized method of claim 1, wherein the actions comprise at least one purchase of an item.

3. The computerized method of claim 1, wherein the actions comprise at least one instruction to add an item to a shopping cart.

4. The computerized method of claim 1, wherein the actions comprise at least one click-through event in which a user clicks on an item.

5. The computerized method of claim 1, wherein the analysis of the prior query-related behavior comprises:

associating result-generating functions with the identified patterns to provide classification results; and

creating the classification mapping information based on the classification results.

6. The computerized method of claim 1, wherein the analysis of the prior query-related behavior operates by weighting the query-related behavior based on time.

7. The computerized method of claim 1, wherein the result-generating functions provide different respective strategies for ranking query results.

8. The computerized method of claim 1, wherein the plurality of result-generating functions are associated with different respective classifications.

9. The computerized method of claim 8, wherein at least one classification indicates that a query is predominately used to identify a title of a desired item.

10. The computerized method of claim 8, wherein at least one classification indicates that a query is predominately used to identify a theme of a desired item.

11. The computerized method of claim 8, wherein at least one classification indicates that a query is predominately used to identify a desired item having a defined brand.

12. The computerized method of claim 8, wherein at least one classification indicates that a query is predominately used to identify a desired item having a defined temporal status.

13. The computerized method of claim 8, wherein at least one of the result-generating functions is used as a default if a classification of the individual user'"'"'s query cannot be determined.

14. The computerized method of claim 8, further comprising determining a classification corresponding to the user'"'"'s query based on the entirety of the query.

15. The computerized method of claim 8, comprising determining a classification corresponding to the user'"'"'s query based on at least one key term in the user'"'"'s query, in the event that classification cannot be performed based on the entirety of the query.

16. The computerized method of claim 1, further comprising:

receiving another query from another user, the another query being substantially similar to the same query;

generating and providing query results to the another user at least in part in response to the receiving of the another query; and

identifying actions taken by the another user upon receiving the query results to determine query-related behavior of the another user after submitting the another query; and

wherein the result-generating function associated with the same query is also based at least in part on the determined query-related behavior of the another user after submitting the another query.

17. The computerized method of claim 16, wherein the result-generating function associated with the same query is also associated with the another query that is substantially similar to the same query.

18. The computerized method of claim 17, further comprising:

receiving the another query that is substantially similar to the same query from a user;

identifying, at least in part in response to the receiving of the another query from the user, the result-generating function; and

generating query results for the user that submitted the another query using the identified result-generating function.

19. An apparatus for providing query results based at least in part on user intent, comprising:

one or more processors;

memory;

a data mining module, stored in the memory and executable on the one or processors to;

collect information regarding behavior of a plurality of users after each user of the plurality of users has submitted a same query; and

based at least in part on the collected information, identify a pattern of behavior of the plurality of users after each of the plurality of users has submitted the same query, the pattern of behavior evincing a predominant intent of the plurality of users submitting the same query; and

a query module, stored in the memory and executable on the one or processors to;

receive the same query from a subsequent user that is different from the plurality of users after the collecting of the information and after the identifying of the pattern of behavior; and

generate search results for the subsequent user at least in part in response to the receiving of the same query from the subsequent user, wherein the search results generated for the subsequent user are tailored to a probable intent of the subsequent user submitting the same query, and wherein the probable intent of the subsequent user is determined based at least in part on the pattern of behavior of the plurality of users that have previously submitted the same query.

20. The apparatus of claim 19, wherein the pattern of behavior indicates that the same query is predominately used to identify a title of a desired item.

21. The apparatus of claim 19, wherein the pattern of behavior indicates that the same query is predominately used to identify a theme of a desired item.

22. The apparatus of claim 19, wherein the pattern of behavior indicates that the same query is predominately used to identify a desired item having a defined brand.

23. The apparatus of claim 19, wherein the pattern of behavior indicates that the same query is predominately used to identify a desired item having a defined temporal status.

24. The apparatus of claim 19, wherein the data mining module is further operative to identify the behavior of the users by identifying actions that the users have taken upon receiving query results.

25. The apparatus of claim 19, wherein the data mining module is further operative to create classification mapping information that maps queries with patterns of behavior predominately associated with the queries.

26. The apparatus of claim 19, wherein the query module provides a plurality of result-generating functions, wherein the result-generating functions apply different respective strategies for producing search results.

27. The apparatus of claim 26, wherein the query module is operative to select one of the plurality of result-generating functions in response to an input of the same query, based at least in part on the pattern of behavior assessed by the data mining module.

28. The apparatus of claim 26, wherein the result-generating functions are operative to provide different respective strategies for ranking query results.

29. The apparatus of claim 28, wherein at least one of the result-generating functions is operative to generate results that give relative priority to result items that have a title-based affiliation with a query.

30. The apparatus of claim 28, wherein at least one of the result-generating functions is operative to generate results that give relative priority to result items that have a theme-based affiliation with a query.

31. A system comprising the apparatus of claim 19, the apparatus being communicatively coupled to a plurality of user devices via a digital network.

32. A computer-implemented method for providing query results based at least in part on user intent, the method comprising:

under control of one or more computer systems configured with executable instructions,

collecting information regarding behavior of a plurality of users after each user of the plurality of users has submitted a same query;

based at least in part on the collected information, identifying a pattern of behavior of the plurality of users after said each user of the plurality of users has submitted the same query, the pattern of behavior evincing a predominant intent of the plurality of users submitting the same query;

receiving the same query from a subsequent user that is different from the plurality of users after the collecting of the information and after the identifying of the pattern of behavior; and

generating search results for the subsequent user at least in part in response to the receiving of the same query from the subsequent user, wherein the search results generated for the subsequent user are tailored to a probable intent of the subsequent user submitting the same query, and wherein the probable intent of the subsequent user is determined based at least in part on the pattern of behavior of the plurality of users that Amended have previously submitted the same query.

33. A computerized method for providing query results based at least in part on user intent, comprising:

receiving, at a computing device, a query from each of a plurality of users, each query of queries received from the plurality of users being a same query;

generating and providing query results to each of the plurality of users at least in part in response to the receiving of the query;

identifying actions taken by each of the plurality of users upon receiving the query results to determine query-related behavior of the plurality of users after submitting the query;

associating a result-generating function with the query based at least in part on the determined query-related behavior of the plurality of users after submitting the query;

receiving the query from an individual user, the query received from the individual user being the same query previously submitted by each of the plurality of users;

identifying, at least in part in response to the receiving of the query from the individual user and using classification mapping information, the result-generating function associated with the query, wherein the classification mapping information maps queries into different result-generating functions based on an analysis of prior query-related behavior of a plurality of users; and

generating query results for the individual user using the identified result-generating function that is based at least in part on the determined query-related behavior of the plurality of users that have submitted the query;

wherein the analysis of the prior query-related behavior operates by weighting the query-related behavior based on time and wherein the analysis comprises;

identifying patterns in the actions that evince a predominant intent of the users in entering the queries;

associating result-generating functions with the identified patterns to provide classification results; and

creating the classification mapping information based on the classification results;

and wherein the different result-generating functions are associated with different respective classifications and at least one of the result-generating functions is used as a default if a classification of the individual user'"'"'s query cannot be determined.