User segment population techniques

• US 8,117,202 B1
• Filed: 12/17/2009
• Issued: 02/14/2012
• Est. Priority Date: 04/14/2005
• Status: Active Grant

• Current Assignee: AudienceScience, Inc.
• Original Assignee: AudienceScience, Inc.
• Inventors: Kozyrczak, Maciek, Nayfeh, Basem, Chen, Cheng Che
• Primary Examiner(s): Le, Debbie
• Application Number:
  US12/640,900
• Time in Patent Office: 789 Days
• Field of Search: None
• US Class Current: 707/736
• CPC Class Codes: G06F 16/90 Details of database functions independent of the retrieved data types
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A facility for identifying users that have an interest in a selected subject is described. For each of a number of actual users, the facility establishes a user behavior record describing the behavior of the actual user. The facility also establishes a user behavior record describing behavior of a virtual user. The behavior described by the user behavior record for the virtual user is made up of actions each tending to indicate an interest in the selected subject. The facility identifies a proper subset of the actual users whose user behavior records are most similar to the user behavior record of the virtual user as interested in the selected subject.

Claims

1. A method in a computing system for identifying users among a plurality of users that have an interest in a selected subject, comprising:

for each of a plurality of actual users, establishing a user behavior record describing behavior of the actual user;

establishing a user behavior record describing behavior of a virtual user, the behavior described by the user behavior record for the virtual user comprising actions each tending to indicate an interest in the selected subject; and

identifying as interested in the selected subject a proper subset of the plurality of actual users whose user behavior records are most similar to the user behavior record established for the virtual user.

2. The method of claim 1, further comprising presenting an advertising message relating to the selected subject to the identified users.

3. The method of claim 1 wherein the behavior described by the user behavior record established for the virtual user comprises web-browsing actions.

4. The method of claim 1 wherein the behavior described by the user behavior record established for the virtual user comprises web page retrieval actions.

5. The method of claim 4, further comprising selecting the actions comprising the behavior described by the user behavior record for the virtual user by searching for web pages that contain query terms relating to the selected subject.

6. The method of claim 1 wherein the identification is performed by applying a cohort matching technique to the user behavior records for the actual users and the virtual user.

7. The method of claim 6 wherein the applied cohort matching technique is item recommendations.

8. The method of claim 6 wherein the applied cohort matching technique is user correlation.

9. The method of claim 6 wherein the applied cohort matching technique is automatic segmentation.

10. The method of claim 6 wherein the applied cohort matching technique is auto-clustering.

11. The method of claim 6 wherein the applied cohort matching technique comprises:

determining, for each of the user behavior records established for an actual user, a distance between (a) the behavior described by user behavior records established for the actual user and (b) the behavior described by user behavior records established for the virtual user; and

identifying users among the plurality of actual users having the smallest distances from the virtual user.

12. One or more computer memories collectively containing a user behavior data structure, comprising:

for each of a plurality of actual users, a user behavior record describing behavior of the actual user; and

a user behavior record describing behavior of a virtual user, the behavior described by the user behavior record for the virtual user comprising actions each tending to indicate an interest in a selected subject,

the contents of the data structure being adapted to use by a cohort matching technique to select the users of the plurality of actual users whose behavior is most similar to that of the virtual user and identify the selected users as interested in the selected subject.

13. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user comprises web-browsing actions.

14. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes text selection actions.

15. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes text reading actions.

16. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes video selection actions.

17. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes video watching actions.

18. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes audio selection actions.

19. The method of claim 12 wherein the behavior described by the user behavior record for the virtual user includes audio listening actions.

20. A computer-readable storage medium whose contents cause a computing system to perform a method for determining a relative ranking of users among a plurality of users indicating the extent to which each user is interested in a selected subject, the method comprising:

among a plurality of available actions, selecting a proper subset of actions whose performance reflects interest in the selected subject;

for each user of the plurality, determining (a) the number of total selected actions performed by the user, (b) the number of distinct selected actions performed by the user, (c) the number of distinct unselected actions performed by the user, and (d) the number of total unselected actions performed by the user; and

determining a relative ranking of the plurality of users that reflects the determination of (a)-(d) for each user.

21. The computer-readable storage medium of claim 20 wherein the determined ranking (1) favors users who performed a higher number of total selected actions;

(2) among users who performed the same number of total selected actions, favors users who performed a higher number of distinct selected actions;

(3) among users who performed the same number of total selected actions and distinct selected actions, favors users who performed a lower number of distinct unselected actions; and

(4) among users who performed the same number of total selected actions, distinct selected actions, and distinct unselected actions, favors users who performed a lower number of total unselected actions.

22. The computer-readable storage medium of claim 20, the method further comprising selecting a proper subset of the plurality of users as being most interested in the selected subject.

23. The computer-readable storage medium of claim 22, the method further comprising presenting an advertising message relating to the selected subject to the identified users.

24. The computer-readable storage medium of claim 20 wherein at least a portion of the selected actions are requesting a particular web page among a plurality of web pages.

25. The computer-readable storage medium of claim 20, the method further comprising:

searching the plurality of web pages to identify those that contain keywords relating to the selected subject, and

for each of the identified web pages, selecting the action of retrieving the identified web page.

26. A computing system having a processor that determines a relative ranking of users among a plurality of users indicating the extent to which each user is interested in a selected subject, comprising:

an action selection system that selects, among a plurality of available actions, a proper subset of actions whose performance reflects interest in the selected subject;

a user behavior analysis subsystem that determines, for each user of the plurality, (a) the number of total selected actions performed by the user, (b) the number of distinct selected actions performed by the user, (c) the number of distinct unselected actions performed by the user, and (d) the number of total unselected actions performed by the user; and

a user ranking subsystem that determines, with the processor, a relative ranking of the plurality of users that reflects the determination of (a)-(d) for each user.

27. The computing system of claim 26 wherein the determined ranking (1) favors users who performed a higher number of total selected actions;

(2) among users who performed the same number of total selected actions, favors users who performed a higher number of distinct selected actions;

(3) among users who performed the same number of total selected actions and distinct selected actions, favors users who performed a lower number of distinct unselected actions; and

(4) among users who performed the same number of total selected actions, distinct selected actions, and distinct unselected actions, favors users who performed a lower number of total unselected actions.