REQ: Bayesian header filter

[Markus]

This request may seem rather outlandish, but it would be nice
if there were some kind of Bayesian header filter available in
Newspro. What I have in mind is a simple filter, much like a normal
E-Mail spam filter, that could be trained by either marking headers
as junk or as valid Posts. After some initial effort from the user, such
a filter would find almost all SPAM/junk postings automatically.
I anticipate that this would make the user experience much more
enjoyable.

Markus

[bassie]

I have spambayes for email. The problem is however that spam mails tend to change their way of spamming. At the end it throws away all of my email.

So i don't know if this works for usenet, maybe it does work.