Combining client and server classifiers to achieve better accuracy and performance results in web page classification

• US 9,223,888 B2
• Filed: 09/08/2011
• Issued: 12/29/2015
• Est. Priority Date: 09/08/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Jiang, Wei, Rovinsky, Vladimir, Hutchings, Bryce
• Primary Examiner(s): ABEDIN, SHANTO
• Application Number:
  US13/228,424
• Publication Number:
  US 20130067590A1
• Time in Patent Office: 1,573 Days
• Field of Search: 713/152, 713/154, 713/168, 726 1- 7, 726 12- 14, 726 22- 30, 386/261, 386/262, 709224-226, 709/229, 725/30, 725/34, 725/35, 725/46, 715/747, 707/748
• US Class Current: 1/1
• CPC Class Codes: G06F 16/958 Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking

Abstract

In one embodiment, an internet monitor service may use a final content rating to determine access to a webpage. A monitor client 102 may generate a client content rating of a webpage 104. The monitor client 102 may factor the client content rating with a server content rating of the webpage 104 to determine a final content rating for the webpage 104.

Claims

1. A machine-implemented method, comprising:

generating using a monitor client a client content rating describing a real-time classification of dynamic webpage content for a webpage;

factoring using the monitor client the client content rating with a server content rating describing static webpage content of the webpage to determine a final content rating for the webpage based on a client confidence score describing certainty in the client content rating and a server confidence score describing certainty in the server content rating; and

storing in a client cache the server content rating as a cached content rating.

2. The method of claim 1, further comprising:

blocking access to the webpage based on the final content rating.

3. The method of claim 1, further comprising:

allowing access to the webpage based on the final content rating.

4. The method of claim 1, further comprising:

receiving the server content rating of the webpage from a monitor server.

5. The method of claim 1, further comprising:

receiving a static server content rating from a monitor server as the server content rating; and

using the static server content rating as the final content rating.

6. The method of claim 1, further comprising:

using the cached content rating as the final content rating.

7. The method of claim 1, further comprising:

ignoring the cached content rating after a client cache validity period has passed.

8. The method of claim 1, further comprising:

executing a real-time classification of at least one of a webpage image and a text content of the webpage to determine the client content rating.

9. The method of claim 1, further comprising:

calculating the client confidence score and the client content rating as part of a single operation.

10. The method of claim 1, further comprising:

generating a telemetry report on the final content rating to be sent to a monitor server.

11. A tangible machine-readable storage device having a set of instructions detailing a method stored thereon that when executed by one or more processors cause the one or more processors to perform the method, the method comprising:

generating using a monitor server a server content rating describing static webpage content for a webpage;

storing the server content rating in a static server content rating storage device; and

sending the server content rating to a monitor client for factoring with a client content rating describing dynamic webpage content of the webpage generated at the monitor client to determine a final content rating for the webpage based on a client confidence score describing certainty in the client content rating and a server confidence score describing certainty in the server content rating.

12. The tangible machine-readable storage device of claim 11, wherein the method further comprises:

receiving a classification request of the webpage from the monitor client.

13. The tangible machine-readable storage device of claim 11, wherein the method further comprises:

maintaining a static server content rating of a known webpage.

14. The tangible machine-readable storage device of claim 13, wherein the method further comprises:

ignoring the static server content rating after a server rating validity period has passed.

15. The tangible machine-readable storage device of claim 11, wherein the method further comprises:

executing a classification of a webpage based on a uniform resource locator.

16. The tangible machine-readable storage device of claim 11, wherein the method further comprises:

calculating the server confidence score and the server content rating as part of a single operation.

17. The tangible machine-readable storage device of claim 11, wherein the method further comprises:

receiving a telemetry report from the monitor client.

18. The tangible machine-readable storage device of claim 17, wherein the method further comprises:

analyzing the telemetry report for adjusting a future server content rating.

19. A computing device, comprising:

a hardware communication interface configured to access a webpage and receives a server content rating with a server confidence score describing certainty in the server content rating;

a hardware processor configured to generate a client content rating of dynamic webpage content for the webpage with a client confidence score describing certainty in the client content rating, to use the server content rating is a static server content rating, and to factor the server content rating with a client content rating of the webpage to determine a final content rating for the webpage when the server confidence score indicates a dynamic server content rating; and

a client cache configured to store the static server content rating as a cached content rating.

20. The computing device of claim 19,

wherein the cached content rating has a client cache validity period.