gnu-emacs-sources
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [off] Quantity of Spam


From: Martin Jost
Subject: Re: [off] Quantity of Spam
Date: Wed, 19 May 2004 12:47:56 +0200

"Richard Stallman" <address@hidden> schrieb im Newsbeitrag news:address@hidden
> Our sysadmins say we already have spam filtering on the mailing list
> and that it does discard many spam messages.  The problem is that
> spammers are learning how to defeat SpamAssassin.  I will suggest
> some other ideas.

Hello,

from my experience with SpamAssassin it is certainly true, that the 
contents-checks get tricked.
I found that RTBL-queries help a bit and using Bayes-filtering with enough 
feedback helps _much_. Bayes needs to be switched on. Then a big chunk of 
ham/spam is needed to train it. Spamassassin uses messages with very large or 
low scores to train automatically. It helps to train by hand for messages which 
are spam, but not "spam enough" to be regarded as such by Spamassassin itself 
(or which are considered as spam but not spam enough to be used for training 
the Bayes filter)

HTH

Martin



reply via email to

[Prev in Thread] Current Thread [Next in Thread]