Re: [SLE] spam

From: "Tage Danielsen" <tage@xxxxxxxxxxxxxxxxxx>

Hello on the list



I am using suse 9.3 with postfixand spamassasing, but i get a lot of spam
anyway.



Does anybody know how to stop mail from .biz?



And I want to stop mail who have an URL for .biz



Thanks in advance

The initial thing to investigate is http://www.rulesemporium.com/.
Select the "rules" tab and read about their various sets of rules.
Pick and choose those which seem to fit your needs best and add them
to /etc/mail/spamassassin/ along with your "local.cf" file. Right
up at the top there is a link to the automatic updater for non-SA
native rule sets, RulesDuJour.

The second thing to investigate is how well your BAYES is going to
work. As is usual for distros these days SUSE has not distributed
the whole SpamAssassin package. You don't have the tools present,
most likely. Some of them are nice. But there is a variant for the
tools "sa-stats.pl" which gathers statistics from the logs that is
better investigating behavior of rule sets. It's hidden away on the
site above at: http://www.rulesemporium.com/programs/sa-stats.txt
Download it and rename it to sa-stats.pl. A quick look comes from
just running it. It'll process the current maillog. If you use
"sa-stats.pl -n 2000 -f maillog*" it will process all maillogs
and give you the top 2000 rules that have gotten hit. The BAYES_99
rule should top out the spam rules that "hit" on your system. And
it should have VERY few hits on ham. (Mine is running 84% of what
gets ultimately marked as spam and 0.04% of what ultimately gets
marked as ham - although I've never noticed a ham that was marked
as spam that included the BAYES_99 markup even though I have stepwise
escallated the score for BAYES_99 up to the spam threshold of 5.0.
I bumped it up and watched for ham with BAYES_99. Then I bumped it
up more. I figured to stop at 5.0, though.)

As it happens I hand train BAYES here because that is practical for
the two people involved. (My partner writes some of the SARE rules.)
In some cases automatic training can actually do better. And in others
it can get off on a wrong foot and be a disaster. I recommend widening
the automatic learn thresholds a little. That helps.

{^_^} Joanne

--
Check the headers for your unsubscription address
For additional commands send e-mail to suse-linux-e-help@xxxxxxxx
Also check the archives at http://lists.suse.com
Please read the FAQs: suse-linux-e-faq@xxxxxxxx

Relevant Pages

  • Re: Test driven development
    ... I don't regularly use TDD yet, and one of the reasons is that in many ... a function spam(), so I write tests for it. ... the function and see that I need to write functions ham() and eggs. ... you extract eggs(). ...
    (comp.lang.python)
  • Bogofilter-0.17.5 - New Current Release
    ... Bogofilter is a mail filter that classifies email as spam or ham ... for spam was described by Paul Graham in his article "A Plan For Spam". ... Bogofilter is run by an MDA script to classify an incoming message as ... Warn if user specified config file doesn't exist. ...
    (comp.os.linux.announce)
  • Bogofilter-0.14.0.1 - New Current Release
    ... Bogofilter is a mail filter that classifies mail as spam or ham ... Updated bogominitrain.pl to work with combined wordlists. ...
    (comp.os.linux.announce)
  • Bogofilter-0.14.1 - New Current Release
    ... Bogofilter is a mail filter that classifies mail as spam or ham ... Updated bogominitrain.pl to work with combined wordlists. ...
    (comp.os.linux.announce)
  • Re: ping: Mara - Apache2 SSL...
    ... I have most of the spam ... ham, and it has to be trained by some intelligent being (i.e. ... It is a bad idea to run sa-learn directly against forwarded ... I think I can tweak the below script to ...
    (alt.2600)