Subject: Multi-layer spam filtering
From: J. M. Mason (mac@orthanc.st.hmc.edu)
Date: Sun May 11 2003 - 15:42:15 AKDT
Have you folks all decided that bogofilter (bayesian, yes?) outperforms
spamassassin in general? My spamassassin install is pretty good; I still
feel like I get way too much spam in my inbox, but then I look in my spam
folder and realize it's about 90% accurate...and it has yet to catalog a
real message as a spam.
As of the most recent release, it's also got a bayesian filter...
I'm curious about this because I may have to try to get a general
spam-filtering solution on a solaris 9 machine with around 300 users.
It seems to me that with the sheer volume of mail being generated, a
bayesian system that worked over everyone's mailbox could be very useful;
of course, there are security issues like crazy with that.
Spamassassin's bayesian can use an sql db to keep track of user prefs and
help with some of that...how does bogofilter handle it? Can it learn
system-wide?
Thoughts?
~Mac~
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
This archive was generated by hypermail 2a23 : Sun May 11 2003 - 15:43:16 AKDT