Adaptive Targeting for Finding Look-Alike Users

• US 20120054040A1
• Filed: 08/30/2010
• Published: 03/01/2012
• Est. Priority Date: 08/30/2010
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Bagherjeiran, Abraham, Tang, Renjie, Ratnaparkhi, Adwait, Zhang, Zengvan, Hatch, Andrew, Parekh, Ralesh
• Granted Patent: US 9,087,332 B2
• Time in Patent Office: Days
• US Class Current: 705/14.66
• CPC Class Codes: G06Q 30/00 Commerce
• CPC Class Codes: G06Q 30/0251 Targeted advertisements
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute

Abstract

A method for adaptive display of internet advertisements to look-alike users using a desired user profile dataset as a seed to machine learning modules. Upon availability of a desired user profile, that user profile is mapped other look-alike users (from a larger database of users). The method proceeds to normalize the desired user profile object, proceeds to normalize known user profile objects, then seeding a machine-learning training model with the normalized desired user profile object. A scoring engine uses the normalized user profiles for matching based on extracted features (i.e. extracted from the normalized user profile objects). Once look-alike users have been identified, the internet display system may serve advertisements to the look-alike users, and analyze look-alike users'"'"' behaviors for storing the predicted similar user profile objects into the desired user profile object dataset, thus adapting to changing user behavior.

Claims

1. A computer-implemented method for adaptive display of an advertisement to look-alike users using a desired user profile dataset, the method comprising:

normalizing at least one desired user profile object, and storing in a computer memory as a normalized desired user profile object;

normalizing at least one known user profile object, and storing in a computer memory as a normalized known user profile object;

seeding a machine-learning training model with the normalized desired user profile object;

matching the normalized desired user profile object to the normalized known user profile object, using an extracted feature from the normalized desired user profile object, and producing a predicted similar user profile object;

serving the advertisement using the predicted similar user profile object; and

adapting to changing user behavior by storing the predicted similar user profile object into the desired user profile object dataset.

2. The method of claim 1, wherein the normalizing comprises selecting only specific features of a user profile.

3. The method of claim 1, wherein the normalizing comprises modifying data items so as to facilitate comparisons.

4. The method of claim 1, wherein the normalizing includes a temporal interpretation of temporal components of the at least one desired user profile object.

5. The method of claim 1, wherein the seeding comprises an advertiser-provided customer list.

6. The method of claim 1, wherein the seeding comprises an advertiser-provided advertisement for monitoring the behavior of users who respond to the advertiser-provided advertisement.

7. The method of claim 1, wherein the matching comprises a machine-learning scoring sub-system.

8. The method of claim 1, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a clustering model.

9. The method of claim 1, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a classifier.

10. An advertising server network for adaptive display of an advertisement to look-alike users using a desired user profile dataset, comprising:

a module for normalizing at least one desired user profile object, and storing in a computer memory as a normalized desired user profile object;

a module for normalizing at least one known user profile object, and storing in a computer memory as a normalized known user profile object;

a module for seeding a machine-learning training model with the normalized desired user profile object;

a module for matching the normalized desired user profile object to the normalized known user profile object, using an extracted feature from the normalized desired user profile object, and producing a predicted similar user profile object;

a module for serving the advertisement using the predicted similar user profile object; and

a module for adapting to changing user behavior by storing the predicted similar user profile object into the desired user profile object dataset.

11. The advertising server network of claim 10, wherein the normalizing comprises selecting only specific features of a user profile.

12. The advertising server network of claim 10, wherein the normalizing comprises modifying data items so as to facilitate comparisons.

13. The advertising server network of claim 10, wherein the seeding comprises an advertiser-provided customer list.

14. The advertising server network of claim 10, wherein the seeding comprises an advertiser-provided advertisement for monitoring the behavior of users who respond to the advertiser-provided advertisement.

15. The advertising server network of claim 10, wherein the matching comprises a machine-learning scoring sub-system.

16. The advertising server network of claim 10, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a linear model.

17. The advertising server network of claim 10, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a clustering model.

18. The advertising server network of claim 10, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a classifier.

19. A computer-implemented method for adaptive display of an advertisement to look-alike users comprising:

seeding a machine-learning training model with at least one desired user profile object;

matching the at least one desired user profile object to at least one known user profile object, the at least one desired user profile object from a desired user profile dataset;

storing one or more predicted similar user profile objects, wherein at least one of the predicted similar user profile objects correlates to the at least one desired user profile object using an extracted feature from the at least one desired user profile object;

serving the advertisement using the at least one of the predicted similar user profile objects; and

adapting to changing user behavior by storing the one or more predicted similar user profile objects into the desired user profile object dataset.

20. The method of claim 19, further comprising:

normalizing at least one of the one or more desired user profile objects.

21. A computer readable medium comprising a set of instructions which, when executed by a computer, cause the computer to perform a display of an advertisement to look-alike users, said instructions for:

normalizing at least one desired user profile object, and storing in a computer memory as a normalized desired user profile object;

normalizing at least one known user profile object, and storing in a computer memory as a normalized known user profile object;

seeding a machine-learning training model with the normalized desired user profile object;

matching the normalized desired user profile object to the normalized known user profile object, using an extracted feature from the normalized desired user profile object, and producing a predicted similar user profile object;

serving the advertisement using the predicted similar user profile object; and

adapting to changing user behavior by storing the predicted similar user profile object into the desired user profile object dataset.

22. The computer readable medium of claim 21, wherein the normalizing comprises selecting only specific features of a user profile.

23. The computer readable medium of claim 21, wherein the normalizing comprises modifying data items so as to facilitate comparisons.

24. The computer readable medium of claim 21, wherein the normalizing includes a temporal interpretation of temporal components of the at least one desired user profile object.

25. The computer readable medium of claim 21, wherein the seeding comprises an advertiser-provided customer list.

26. The computer readable medium of claim 21, wherein the seeding comprises an advertiser-provided advertisement for monitoring the behavior of users who respond to the advertiser-provided advertisement.

27. The computer readable medium of claim 21, wherein the matching comprises a machine-learning scoring sub-system.

28. The computer readable medium of claim 21, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a clustering model.

29. The computer readable medium of claim 21, wherein the matching comprises a machine-learning scoring sub-system for scoring based on a classifier.