On Thu, 2003-01-16 at 17:20, Marco Tabini wrote: > Hello All, > > I've written a statistical spam filter for Evolution. It works by > analyzing the e-mail that you consider spam and eventually learning to > automatically filter it into a special folder. Because it doesn't use a > set of "rigid" rules, it should theoretically adapt dynamically, without > requiring any future reinstall or updates.
What do you use? Bayesian filter?
Also, from your mail it's not clear: is spam discarded? I'd advice
*strongly* against it - false positives do happen. Always, sooner or
later.
> Because I'll be distributing binaries only, I need someone who is using
> Linux. The filter makes use of labels, so you should not be using those
> either.
Huh!? Ok, forget it.
For those who can easily install software on the computer receiving
their mail: use procmail and spamassassin and bogofilter (both are free
Software).
My setup:
- mail comes in
- mail goes through bogofilter
- mail goes through spamassassin
a spamassassin rule gives a score if bogofilter said it's spam.
- mail goes through bogofilter in learning mode
trains bogofilter according to the opinion of spamassassin
- mail goes through procmail filtering, sorting mail into folders.
spamassassin 2.5 will make bogofilter obsolete, I hope.
cheers
-- vbi
--
this email is protected by a digital signature: http://fortytwo.ch/gpg
signature.asc
Description: This is a digitally signed message part
