[wellylug] SpamAssassin/Virus checking 'before' Exchange Server

Brenda O'Hagan brenda at wallace.net.nz
Mon Apr 5 14:56:50 NZST 2004


I've been using "bogofilter" for a couple weeks now.

it's completely bayesian, and is faster and so far it's much more accurate 
than spamassassin . It's catching spam that spamassassin doesn't, and i 
haven't had any false positives.

your results will depends on the quality and quantity of the spam/ham samples 
you feed it

oh, and it's available on mandrake's urpmi
:-)

-Brenda

On Mon, 05 Apr 2004 07:58, Glen Ogilvie wrote:
> Hi Cliff,
>
> I have not setup Bayesian filtering, however, my understanding is that you
> need to collect a large quanitity of both valid recevied email for your
> company and around the same quantity of spam.  This is then given to the
> filter that is used to make it's rules about what it spam.
>
> MailWatch, a http front end to Spamassassin also has a feature to mark spam
> as not spam, improving the filter, after you have set it up.
>
> Glen
>
> On Friday 02 Apr 2004 8:55 pm, Enkidu wrote:
> > Hi Glen,
> >
> > I use Exim, Amavis, Spamassassin, ClamAV. I'm intertested in the
> > Bayesian Filter systems which are supposed to learn. How do you teach
> > it?
> >
> > Cheers,
> >
> > Cliff




More information about the wellylug mailing list