Adaptive junk message filtering system

• US 7,249,162 B2
• Filed: 02/25/2003
• Issued: 07/24/2007
• Est. Priority Date: 02/25/2003
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Platt, John C., Rounthwaite, Robert L., Goodman, Joshua T., Heckerman, David E., Kadie, Carl M.
• Primary Examiner(s): Cardone; Jason
• Assistant Examiner(s): NGUYEN, MINH CHAU
• Application Number:
  US10/374,005
• Publication Number:
  US 20040167964A1
• Time in Patent Office: 1,610 Days
• Field of Search: 709/206, 709/203
• US Class Current: 709/206
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]
• CPC Class Codes: H04L 51/212 using filtering or selective blocking

Abstract

The invention relates to a system for filtering messages—the system includes a seed filter having associated therewith a false positive rate and a false negative rate. A new filter is also provided for filtering the messages, the new filter is evaluated according to the false positive rate and the false negative rate of the seed filter, the data used to determine the false positive rate and the false negative rate of the seed filter are utilized to determine a new false positive rate and a new false negative rate of the new filter as a function of threshold. The new filter is employed in lieu of the seed filter if a threshold exists for the new filter such that the new false positive rate and new false negative rate are together considered better than the false positive and the false negative rate of the seed filter.

Claims

1. A data filtering system, comprising:

a first filter configured to tag messages as junk based at least in part on junk information associated with the messages, the first filter having associated therewith a false positive rate and a false negative rate;

one or more second filters configured to tag the messages as junk based at least in part on junk information associated with the messages, the one or more second filters initially associated with the false positive rate and the false negative rate of the first filter

a filter output configured to receive tagged and untagged messages from the first filter and the one or more second filters;

a user correction component configured to receive user actions relating to the tagged and untagged messages sent to the filter output and to output false positive data and false negative data based on the user actions relating to the tagged and untagged messages sent to the filter output; and

a filter control configured to;

receive the false positive data and the false negative data;

adjust the false positive rate or the false negative rate or both of at least one of the one or more second filters based on its false positive data or its false negative data or both; and

route subsequently received messages between the first filter and the one or more second filters according to a threshold and their respective false positive rates, false negative rates or both.

2. The system of claim 1, the user actions comprising implicitly tagging the messages.

3. The system of claim 1, the user actions comprising explicitly tagging the messages as junk.

4. The system of claim 1, the first filter and the one or more second filters further configured to tag the messages as junk based at least in part on good data.

5. The system of claim 1, wherein the filter control is further configured to adjust the false positive rate and the false negative rate of at least one of the one or more second filters based on the content of other user'"'"'s messages.

6. The system of claim 1, wherein configured to route subsequently received messages between the first filter and the one or more second filters according to a threshold and their respective false positive rates, false negative rates or both comprises configured to route the messages to a filter having a best false positive rate.

7. The system of claim 1, wherein the filter control is further configured to adjust the false positive rate, false negative rate or both of at least one of the one or more second filters after a predetermined number of messages are tagged, a predetermined time has occurred or both.

8. The system of claim 1, wherein the system is further configured to select the threshold from a plurality of generated threshold values comprising an average threshold value over eligible threshold values, a threshold value with lowest false positive rate, or a threshold value that maximizes user'"'"'s expected utility based upon a p* utility function or combinations thereof.

9. The system of claim 1, wherein the system is further configured to select the threshold from a plurality of threshold values.

10. A portable computing device comprising the system of claim 1.

11. The device of claim 10 being a personal data assistant, a telephone or a laptop computer.

16. A computer-readable medium having computer-executable components stored thereon to effect the system of claim 1.

17. A computer comprising the system of claim 1.

18. A network comprising the system of claim 1.

12. A method of facilitating data filtering, comprising:

automatically filtering incoming messages according to a false positive rate and a false negative rate of a seed filter;

receiving user-correction data relating to at least one filtered message;

determining an accuracy of the seed filter based on the user-correction data relating to the at least one filtered message;

training a new filter using the user-correction data;

determining a false positive rate and a false negative rate of the new filter;

determining an accuracy of the new filter based on the false positive rate and the false negative rate of the new filter; and

employing the new filter in lieu of the seed filter if the accuracy of the new filter is better than that of the seed filter.

13. The method of claim 12, the user-correction data comprising user-correction data relating to overriding an initial classification of the at least one filtered message.

14. The method of claim 12, wherein determining the false positive rate and the false negative rate of the new filter comprises determining the false positive rate and the false negative rate of the new filter after a predetermined number of junk and non-junk messages are labeled by a user, a predetermined time has occurred or both.

15. The method of claim 12 implemented by computer-executable instructions stored on one or more computer-readable media.

19. A data filtering system, comprising:

first means for filtering messages, the first means for filtering messages having associated therewith a false positive rate and a false negative rate;

new means for filtering the messages, the new means for filtering the messages trained according to the false positive rate and the false negative rate associated with the first means for filtering the messages;

means for determining a new false positive rate and a new false negative rate associated with the new means for filtering the messages as a function of threshold;

means for determining a threshold of the new means for filtering the messages;

means for employing the new means for filtering the messages in lieu of the first means for filtering the messages if a threshold exists for the new means for filtering the messages such that the new false positive rate and new false negative rate associated with the new means for filtering the messages are together considerd better than the false positive rate and the false negative rate associated with the first means for filtering the messages.