CHARACTER DIFFERENTIATION BASED ON PATTERN RECOGNITION

• US 20100241745A1
• Filed: 03/17/2010
• Published: 09/23/2010
• Est. Priority Date: 03/17/2009
• Status: Active Grant

• Current Assignee: Alcatel-Lucent USA, Inc. (Nokia Corporation)
• Original Assignee: Kindsight Incorporated
• Inventors: Offen, Russell, Edmison, Kelvin, Gaudet, Rob
• Granted Patent: US 8,150,974 B2
• Time in Patent Office: Days
• US Class Current: 709/224
• CPC Class Codes: G06Q 30/0251 Targeted advertisements

Abstract

A method and apparatus for character differentiation based on pattern recognition is provided. A sequence of events is received that have been generated from a given session or connection and determines if the sequence of events is likely to have come from an existing, known character, or if the events are from a previously unknown character. A character identifier (ID) is tagged onto (i.e. associated with) the event for further processing in another system that can, for example, provide different content based on the identified character and an associated preference profile.

Claims

1. A method of character differentiation of on-line communications originating from a subscriber access point coupled to a network, the method comprising:

receiving user generated events from one of a plurality of computing devices coupled to the access point, each event associated with a subscriber identification (ID) and a session identification (ID), the events generated by a communication session between one of the plurality of computing devices and a communication endpoint through the network;

generating a session fingerprint using a plurality of received events for the associated subscriber ID and session ID wherein the session fingerprint comprises a feature vector based upon data contained in the received events; and

storing a character record comprising the generated fingerprint, an associated character identifier (ID) and the subscriber ID.

2. The method of claim 1 wherein the feature vector comprises tokens, each token indexed to a term defined in vocabulary list, the tokens determined based upon communication endpoint location information provided by the received events.

3. The method of claim 2 wherein each token further comprises a weight associated with token defined in the feature vector, the weight defining the relative occurrence of the token in the received events.

4. The method of claim 3 wherein the tokens are defined by host names identified in the event, the tokens form the vector that represent the host name.

5. The method of claim 1 wherein the features include a vocabulary of words commonly present in website host names, as well as time-of-day and major category information.

6. The method of claim 2 wherein the feature vector features selected form the list comprising:

a channel change events, time-of-day events, and metadata information including program title names, descriptions and categories.

7. The method of claim 1 further comprising determining if the character record is active within a defined time interval, wherein if character record is active, the character record is updated with the received events.

8. The method of claim 1 wherein prior to storing the character record, the method further comprising comparing the session fingerprint to one or more stored fingerprints, each associated with a character record sharing a common subscriber ID, to determine if a character fingerprint has already been identified.

9. The method of claim 8 wherein comparing the fingerprints is performed by a matching algorithm to provide a similarity indicator between the determined fingerprint and the stored fingerprints, the matching algorithm providing a value between 0 and 1.

10. The method of claim 9 wherein if the similarity is greater than 0.5 a match is declared and the feature vectors are updated based upon the received events.

11. The method of claim 10 wherein the matching algorithm is a cosine similarity algorithm.

12. The method of claim 1 determining if the received events are associated with an existing active session is further defined based upon an active time interval in which associated events must be received to maintain character record as active.

13. The method of claim 1 wherein the comparing the fingerprint further comprises comparing a time-of-day or day of the week associated with the character record.

14. The method of claim 13 further comprising comparing a classification of websites that were visited.

15. The method of claim 1 wherein the events define a HTTP request portion between the computing device and communication endpoint.

16. The method of claim 1 wherein the session ID is assigned by OS fingerprinting technique to distinguish active communication sessions originating from the same access point.

17. The method of claim 1 wherein the access point is selected from the group comprising:

a digital subscriber line modem;

a Data Over Cable Service Interface Specification (DOCSIS), and a wireless modem.

18. The method of claim 1 wherein the character ID, subscriber identifier and received events are used to determine an associated content preference profile used from determining applicable content to be provided to a user associated with the session based upon user content preferences.

19. A character identification apparatus for character differentiation of on-line communications originating from a subscriber access point coupled to a network, the module comprising:

a processor;

a memory providing instructions for execution by the processor, the instructions for;

receiving user generated events from one of a plurality of computing devices coupled to the access point, each event associated with a subscriber identification (ID) and a session identification (ID), the events generated by a communication session between one of the plurality of computing devices and a communication endpoint through the network;

generating a session fingerprint using a plurality of received events for the associated subscriber ID and session ID wherein the session fingerprint comprises a feature vector based upon data contained in the received events; and

storing a character record comprising the generated fingerprint, an associated character identifier (ID) and the subscriber ID.

20. A computer readable memory storing instructions for execution by a processor, the instructions for performing character differentiation of on-line communications originating from a subscriber access point coupled to a network, the instructions comprising:

receiving user generated events from one of a plurality of computing devices coupled to the access point, each event associated with a subscriber identification (ID) and a session identification (ID), the events generated by a communication session between one of the plurality of computing devices and a communication endpoint through the network;

generating a session fingerprint using a plurality of received events for the associated subscriber ID and session ID wherein the session fingerprint comprises a feature vector based upon data contained in the received events; and

storing a character record comprising the generated fingerprint, an associated character identifier (ID) and the subscriber ID.