On Sonntag, 23. April 2006 04:02 Gaute Lund wrote: > So, I was hoping to get a different opinion here.
I use bayes per server, not per user or domain. I've set autolearn, with everything 8+ points as spam, below +1 as ham. bayes_auto_learn_threshold_spam 8.00 bayes_auto_learn_threshold_nonspam 1.0 In addition, there are several honeypot SPAM harvesters, which are explicitly excluded from scanning via amavis, which are semi-auto fed to SA and learned as SPAM. This is working good for me, although I don't have stats telling me explicitly when bayes made a difference between being tagged or not. I'd like to do this, but didn't find a good tool for that until now. I can say that I get near-zero SPAM, but I have many more tools than SA alone, for example greylisting and such. mfg zmi -- // Michael Monnerie, Ing.BSc ----- http://it-management.at // Tel: 0660/4156531 .network.your.ideas. // PGP Key: "lynx -source http://zmi.at/zmi3.asc | gpg --import" // Fingerprint: 44A3 C1EC B71E C71A B4C2 9AA6 C818 847C 55CB A4EE // Keyserver: www.keyserver.net Key-ID: 0x55CBA4EE
pgp8qR6fYo2hO.pgp
Description: PGP signature
