You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Pitmaster <in...@dszdekooi.nl> on 2006/08/19 11:26:25 UTC

howto let SA find more spam

Hi there,
I am a total newbie. Have SA installed and done nothing sinds.
SA is working and catches +- 5% of the spam. How can I simply help SA to
catch more and keep the speed up?

Greeting and thank's in advance,
Nico
-- 
View this message in context: http://www.nabble.com/howto-let-SA-find-more-spam-tf2131550.html#a5882863
Sent from the SpamAssassin - Users forum at Nabble.com.


Re: howto let SA find more spam

Posted by Matthias Haegele <mh...@linuxrocks.dyndns.org>.
Pitmaster schrieb:
> Hi there,

Hello!.

> I am a total newbie. Have SA installed and done nothing sinds.
> SA is working and catches +- 5% of the spam. How can I simply help SA to
> catch more and keep the speed up?
> 

You could read through:
http://wiki.apache.org/spamassassin/
(use google, searching for: "wiki spamassassin" or "howto spamassassin")

on your local system read:
/usr/share/doc/spamassassin/*
man spamassassin

perhaps you want to use sa-learn to train the bayes-system (needed 200 
ham and spam mails to activate).

use sa-update
upgrade to newer SA version?

> Greeting and thank's in advance,
> Nico

hth
MH

Re: howto let SA find more spam

Posted by mouss <us...@free.fr>.
Pitmaster wrote:
> Hi there,
> I am a total newbie. Have SA installed and done nothing sinds.
> SA is working and catches +- 5% of the spam. How can I simply help SA to
> catch more and keep the speed up?
>
> Greeting and thank's in advance,
> Nico
>   

if you want SA to find more spam, give it more spam:)

more seriously, that depends on your situation. if you have much 
legitimate mail that resembles spam mail, you'll have a hard time. 
otherwise,
follow what Bowie said, but what Maththias said seems very important 
here. I'v tested SA without Bayes for some months (and without upgrading 
rules of course). the difference is important. so train your bayes.