Hello Kenneth, Friday, January 9, 2004, 6:22:41 AM, you wrote:
KP> --On Friday, January 09, 2004 1:53 AM -0700 Bob Proulx <[EMAIL PROTECTED]> KP> wrote: >> For me personally SA is still tagging the spam at a very good rate. I >> am only seeing these types of spams in my caughtspam folder. KP> I'm seeing them in my Inbox, with BAYES_00 overriding other rules that KP> would easily push the spam over the threshold. I just scored BAYES_00 to KP> zero to stop that, until some more permanent measure is available. To the best of my knowledge, if you're getting Bayes_00 on spam, your bayes database is poisoned. My recommendations: 1) Change the minimum auto-learn threshold. Drop it so low that spam can't sneak in as ham. I run mine at -2. 2) Delete your bayes database and start over. Starting over is easy if you can do an immediate sa-learn on 200+ each or 500+ each GUARANTEED spam and ham. If you don't have that much email saved, then sa-learn (correctly!) whatever you can as it comes in. 3) ALWAYS sa-learn low-bayes spam. IMO you should sa-learn all email possible, but mis-learned or not-recognized spam and ham are the most important emails to sa-learn. Bob Menschel ------------------------------------------------------- This SF.net email is sponsored by: Perforce Software. Perforce is the Fast Software Configuration Management System offering advanced branching capabilities and atomic changes on 50+ platforms. Free Eval! http://www.perforce.com/perforce/loadprog.html _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
