2006-02-01
Abstract
A common criticism of statistical spam filters (such as SpamBayes, DSPAM and POPFile) is that they can be 'poisoned' by inserting random words into spam messages. John Graham-Cumming looks at the evidence and attempts to answer the question: does Bayesian poisoning exist?
Copyright © 2006 Virus Bulletin
The full article is available to registered users. Click here for free registration or, if you already are a registered user, login to access the full article.
Poll
Do you use the same password(s) across multiple websites?Leave a comment
View 4 comments

VB100 certification
This month VB's test team put 26 products to the test on
Windows Server 2008. John Hawes has the full results.
See full results.
Virus Bulletin currently has 190,944 registered users.

