spamassassin and sa-learn

syl20 · 04-27-2005, 02:39 AM

Hi,
I use SA for 4 months now, on a mail server with one hundred pop accounts.
When a spam pass through SA, I ask users to return mail at spam@my-gw.

I would like to use sa-learn to make a better spam filtering, but when users forward spam at sapm@my-gw, a new header is added and the text begin by > on each line. How could I do to use sa-learn --spam better?

Thx

bigrigdriver · 04-27-2005, 07:44 PM

Perhaps this article will help.

syl20 · 04-28-2005, 03:50 AM

Hi bigrigdriver,
thx for the article.
But I'm not sure that I can use forwarded emails to learn from them as spam...
I can't read the hundreds of emails which pass though my mail server to see if they are spam or no!!

How do SA users to make bayesian filtering better?