Peter Teunissen wrote: > > > OTH, I just looked at some of my recent image spam and when I substract > the bayes score from the final score, most of it would still be tagged > as spam (I use a threshold of 5).
True, but those messages that get through tend to have almost no score assigned to them. :-( That said, if I keep feeding the false negatives into the bayes database then it does a pretty good job. On my IMAP server at home I have a set of scripts that harvest a particular folder in the account and feed all messages to sa-learn. If a spam slips through then I just move it into that folder and forget about it. :-) -- George Borisov DXSolutions Ltd
signature.asc
Description: OpenPGP digital signature