PROFILE PREDICTION FOR SHARED COMPUTERS

• US 20120109956A1
• Filed: 10/29/2010
• Published: 05/03/2012
• Est. Priority Date: 10/29/2010
• Status: Active Grant

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Ramaiyer, Viswanathan, Verma, Amit Kumar, Prakash, Hastagiri
• Granted Patent: US 9,536,000 B2
• Time in Patent Office: Days
• US Class Current: 707/737
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 30/0277 Online advertisement

Abstract

A profile prediction system may identify one of multiple user profiles for a single computer. For example, a home computer may have multiple users that may not be targeted unless the user on the home computer can be identified. The system'"'"'s user identification may be based on a clustering model that considers various browsing characteristics to identify different clusters that each correspond to a particular user or user profile. The model may be generated and refined by tracking web browsing over multiple sessions. Future activity on the computer may be used to identify which user is the source of the activity and the user may receive targeted content including advertisements.

Claims

1. A method for generating a profile prediction model comprising:

identifying sessions of behavior from a single source;

analyzing each of the identified sessions to cluster them based on patterns, wherein each of the identified sessions is added to one of a plurality of clusters;

developing a user profile for each of the plurality of clusters, wherein the user profile is based on the behavior from the one or more sessions that are part of the cluster for that user profile;

generating a model for analyzing behavior to associate that behavior with one of the clusters;

identifying the user profile based on the model, wherein monitored behavior for a particular session is tracked and the model identifies the user profile based on the monitored behavior; and

targeting content based on the identified user profile.

2. The method of claim 1 wherein the user profile comprises key characteristics for one or more sessions associated with the user profile.

3. The method of claim 2 wherein the key characteristics comprise at least one of a beginning site, a most frequented site, a category of site most frequented, and a time of day.

4. The method of claim 1 wherein the content that is targeted comprises advertisements that are displayed based on the identified user profile.

5. The method of claim 1 wherein the single source comprises a computer accessed by multiple users, wherein the multiple users are each associated with one of the user profiles.

6. In a computer readable medium having stored therein data representing instructions executable by a programmed processor for identifying a user profile from a single source with multiple user profiles, the storage medium comprising instructions operative for:

identifying clusters of behavior from a single source;

analyzing each of the clusters;

developing a profile for each of the clusters based on the analysis of the clusters;

monitoring behavior at the single source to identify one of the clusters of behavior, wherein the identified cluster is associated with a developed profile for that identified cluster; and

targeting content based on the developed profile for that identified cluster.

7. The computer readable medium of claim 6 wherein the single source comprises a computer.

8. A system for identifying a user on a shared computer comprising:

a web server coupled with a network and configured to receive requests from a source and return a page to the source in response to each of the requests;

an ad server coupled with the network and configured to provide advertisements with the pages returned in response to the requests;

a profile predictor coupled with the web server and the ad server and configured to analyze the requests, the profile predictor comprising;

a cluster generator that analyzes sessions of the requests and assigns each of the sessions to one of a plurality of clusters;

a profile generator that determines characteristics for each of the clusters and generates a profile for each of the clusters that includes the determined characteristics for that cluster; and

a model generator that produces a model for associating a profile with requests received at the web server.

9. The system of claim 8 wherein the source comprises a user device.

10. The system of claim 9 wherein the user device comprises a computer that is accessible by multiple users and the multiple users are each associated with one of the generated profiles.

11. The system of claim 10 wherein the model identifies one of the multiple users based on requests received at the web server.

12. The system of claim 8 wherein the ad server provides targeted advertisements in response to received requests that are targeted based on identifying a profile for those requests.

13. The system of claim 8 wherein response from the web server to the requests utilizes the model to target content for those requests based on profile associated with those requests.

14. The system of claim 8 wherein the generated profiles each comprise characteristics including a first site visited, a second site visited, a most frequent site, a site with most time spent, site categories browsed, and/or time of day.

15. The system of claim 14 wherein the characteristics of the requests received at the web server are analyzed with the model for matching with one of the profiles.

16. The system of claim 8 wherein the generated profiles each comprise behavior data.

17. The system of claim 16 wherein the behavior data comprises browsing history, click history, or multimedia usage.

18. A method for identifying and predicting users of a computer, the method comprising:

generating a model that identifies the users, the generating comprising;

analyzing a plurality of sessions from the computer;

clustering the sessions based on common characteristics;

generating a profile for each of the clusters, wherein each of the profiles is associated with one of the users; and

producing a model that associates common characteristics with the profile;

identifying an active user from the users, the identifying comprising;

receiving requests from the computer;

analyzing, with the model, the common characteristics of the requests; and

predicting the user submitting the requests from the computer based on the analysis.

19. The method of claim 18 further comprising:

targeting the predicted user based on the profile associated with that user, wherein the targeting comprises targeted advertisements displayed in response to the requests.

20. The method of claim 18 wherein the common characteristic comprise browsing habits, further wherein the browsing habits include which sites are visited, how long sites are visited, how frequently sites are visited, and/or categories of visited sites.