on Wed, 30 Apr 2003 10:18:14AM +0200, Christian Jaeger insinuated:
> At 0:00 Uhr -0700 30.04.2003, nate wrote:
> >I personally get maybe 1 in 1,000 false postives, though some users
> >that have accounts on my machines apparently get much higher rates
> >so I had to add them to the white list. I don't know why, I've been
> >using it since september and have never had a bad false
> >positive(all false positives have been mailing list postings, which
> >are hardly critical, unlike say an e-commerce recipt or some
> >personal email like a job offer :) )
>
> SpamAssassin 2.53 is significantly better than the version in woody.
> Especially once the "bayes" filter begins its work (after at least
> 200 spams have been reported).
Cool, I just upgraded to 2.53, and it seems better. Can someone
explain to me how a Bayesian filter would work in a static context?
Does it train itself based on the threshold you provide? My
impression (though this could be wrong, as I've only ever used
spamassassin) is that on other commonly-used Bayesian spamfilters, you
have to manually train it on x number of emails before it learns what
you consider spam and what you don't. How does spamassassin -- which
is procmail-based -- train itself?
Curious,
</nori>
--
.~. nori @ sccs.swarthmore.edu
/V\ http://www.sccs.swarthmore.edu/~nori/jnl/
// \\ @ maenad.net
/( )\ www.maenad.net
^`~'^
get my (*new*) key here:
http://www.maenad.net/geek/gpg/7ede5499.asc
(please *remove* old key 11e031f1!)
Attachment:
pgpTIwVnjMavo.pgp
Description: PGP signature