Trees of classifiers for detecting email spam

• US 20070038705A1
• Filed: 07/29/2005
• Published: 02/15/2007
• Est. Priority Date: 07/29/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Rounthwaite, Robert, Heckerman, David, Goodman, Joshua, Chickering, David, Hulten, Geoffrey, Meek, Christopher
• Granted Patent: US 7,930,353 B2
• Time in Patent Office: Days
• US Class Current: 709/206
• CPC Class Codes: H04L 51/212 using filtering or selective blocking

Abstract

Decision trees populated with classifier models are leveraged to provide enhanced spam detection utilizing separate email classifiers for each feature of an email. This provides a higher probability of spam detection through tailoring of each classifier model to facilitate in more accurately determining spam on a feature-by-feature basis. Classifiers can be constructed based on linear models such as, for example, logistic-regression models and/or support vector machines (SVM) and the like. The classifiers can also be constructed based on decision trees. “Compound features” based on internal and/or external nodes of a decision tree can be utilized to provide linear classifier models as well. Smoothing of the spam detection results can be achieved by utilizing classifier models from other nodes within the decision tree if training data is sparse. This forms a base model for branches of a decision tree that may not have received substantial training data.

Claims

1. A system that facilitates classification of electronic mail, comprising:

a feature detection component that detects at least one feature of at least one obtained email; and

a decision tree classification component that classifies the obtained email based on, at least in part, the features of the obtained email and at least one feature classifier employed via a decision tree.

2. The system of claim 1, the decision tree classification component determines whether an email is spam.

3. The system of claim 1, the decision tree classification component utilizes at least one linear classifier.

4. The system of claim 2, the linear classifier comprising a logistic-regression classifier.

5. The system of claim 2, the linear classifier comprising a support vector machine (SVM).

6. The system of claim 1, the decision tree classification component utilizes at least one classifier constructed from a decision tree.

7. The system of claim 1, the decision tree classification component smoothes at least one classification result via employment of at least one classifier that is based on, at least in part, at least one other classifier utilized in at least one other node of the decision tree.

8. The system of claim 1, the decision tree classification component employs a linear classifier constructed from compound features based on, at least in part, nodes of the decision tree.

9. The system of claim 8, the compound features based on, at least in part, at least one internal node of the decision tree.

20. A device employing the system of claim 1 comprising at least one selected from the group consisting of a computer, a server, and a handheld electronic device.

10. A method for facilitating classification of electronic mail, comprising:

obtaining an email from a source; and

utilizing a decision tree of classifier models to detect whether the email is spam.

11. The method of claim 10 further comprising:

employing a linear model as at least one classifier of the decision tree.

12. The method of claim 11, the linear model comprising a logistic-regression model.

13. The method of claim 11, the linear model comprising a support vector machine (SVM).

14. The method of claim 10 further comprising:

employing a decision tree as at least one classifier of the decision tree.

15. The method of claim 10 further comprising:

smoothing at least one detection result utilizing a classifier model that incorporates, at least in part, a classifier model built for another node of the decision tree.

16. The method of claim 10 further comprising:

employing at least one linear model constructed from compound features that are based on, at least in part, nodes of the decision tree.

17. The method of claim 16 further comprising:

constructing compound features based on, at least in part, at least one internal node of the decision tree.

19. A device employing the method of claim 10 comprising at least one selected from the group consisting of a computer, a server, and a handheld electronic device.

18. A system that facilitates classification of electronic mail, comprising:

means for detecting features of at least one obtained email; and

means for classifying the obtained email based on, at least in part, the features of the obtained email and at least one feature classifier employed via a decision tree.