Simon Byrnand said: > > - for spam, must have 3 head hits and 3 body hits > > Why ? This seems a bit arbitrary to me. Either we trust the scoring or we > don't :) What is magic about 3 in particular ?
Yeah, not sure myself. if I recall correctly it gave a stronger statistical basis to ensure we avoided FPs. (tested by Ed Allen Smith) > >add this one: > > > > - previous hits must be < bayes_auto_learn_threshold_nonspam or > > > bayes_auto_learn_threshold_spam > > Hmm, well for starters that would prevent whitelisted spam (such as spam on > this list when people have the list whitelisted) from being autolearnt, and > depending on your point of view that could be a good thing or a bad thing.... Good thing ;) Auto-learning has to be totally trustworthy - *no* FPs or FNs as much as possible. If we miss a bit of spam/ham, c'est la vie -- there's always plenty more and it'll get enough training, eventually. More important is that it has to work unsupervised. (Supervised auto-learning BTW, as I posted, is easy and can be implemented in the procmailrc. Hence SpamAssassin's auto-learning's emphasis on the _unsupervised_ variety.) --j. ------------------------------------------------------- This SF.Net email is sponsored by: INetU Attention Web Developers & Consultants: Become An INetU Hosting Partner. Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission! INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
