Spamassassin, Bayesian Filtering

[Bernhard Reiter]

Am Dienstag, 29. November 2005 16:57 schrieb Markus F. Kraemer:
> after a little brake I gave Bayesian filtering another shot yesterday
> morning.
>
> After deleting the /kolab/var/amavisd/.spamassassin folder I started
> over with sa-learn and voila it worked.

Great!

> I have no idea what went wrong except that I was feeding sa-learn
> several huge mbox-style folders(>800mb) with spam mails over night.
> Could it be that I was feeding to much spam into sa-learn and the
> corresponding db´s??
>
> Whatever, it is working now.
>
> Shall I add a Step 7 for debuging in the "Fighting spam"-section?

Sure, why not? :)