[LRUG] Anti-spam measures

[Paul Battley]

2008/11/27 Paul Dix <paul at pauldix.net>:
> This article outlines a really good approach. It catches a lot and
> I'd say the best first line is to follow the steps in that first post.
> If you're still having spam problems then break out the big guns with
> naive bayes.

Bayesian classification originally worked quite well for me, but the
spammers are getting smarter with their content. By mid-way through
this year I was getting too many false negatives and false positives.

The simple solutions are the best.

P.