Hello,
I have question and hope that somebody will help me. Please...
So, few days ago I installed and configured SpamAssassin on Linux. The SA
is works fine, but he is catching only about 20% of spam emails. I believe
that I should configure Bayesian with SA. Is it correct? How can I check
that I really have Bayesian (I have most recently version of SA and should
have Bayesian, but I want to double check it)? How can I configure and start
Bayesian? How can I use "auto-learn" function and what should I do for it?
The bayes engine will kick in by default once it has learned enough spam and non-spam messages.. doing it by auto-learn can take a while, but you can expedite the process by using sa-learn to hand-feed the bayes engine.
This is probably a good idea anyway, as you'll be able to add in "low scoring" spam emails that are designed with mis-spelled words to get around the rules. things like rem0ve make great bayes tokens for spam :) Auto-learn only learns when the score of an email is very high or very low so it won't likely feed any of these kinds of messages into the learner..
if you fire the command-line spamassassin -D --lint one of the things the debug output will spit out is if it has a reason not to use bayes.
ie: debug: debug: Only 1 spam(s) in Bayes DB < 200
(SA needs at least 200 spam and 200 nonspams in the bayes DB to use bayes.. it also has some critera for the number of tokens in the database, so some people need to feed it more)
------------------------------------------------------- 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
