Hello Clive, Wednesday, December 24, 2003, 8:09:42 AM, you wrote:
CD> I am receiving several spam messages daily in which the message body appears CD> to consist entirely of random words. Check again, and I think you'll find that the message has nothing to do with those random words. The message is either in a graphic attachment or an HTML attachment, and the random words are to help the email slip through filters. CD> Spamassassin is not catching these messages. The Bayesian filter has not yet CD> kicked in but I am running uncaught spam through sa-learn. I am concerned CD> about whether the Bayesian filter will be biased by an accumulation of random CD> words and so I am seeking advice as to whether I should continue to run this CD> class of spam through sa-learn. Continuing running it through sa-learn. I used to avoid this, but changed my mind about 2 months back, and have seen no problems learning these random words. What we're really teaching SA about are the headers and the other useful attributes of the message; the random words are just noise which at this level doesn't seem to affect anything (except maybe make the Bayes database grow a little in size). Bob Menschel ------------------------------------------------------- This SF.net email is sponsored by: IBM Linux Tutorials. Become an expert in LINUX or just sharpen your skills. Sign up for IBM's Free Linux Tutorials. Learn everything from the bash shell to sys admin. Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
