spam from the list

Dimitrios sehh at altered.com
Wed Jun 4 12:25:38 BST 2003


On Wed, 04 Jun 2003 11:29:00 +0100 Pete French <pete at twisted.org.uk> wrote:

> is there an online corpus of spam anywhere that could be used for this ?
> I guess it also needs an equivalent quantitiy of 'ggod'email to balance it
> out doesnt it ?

I dont know of any place, but i can send you my 4000 spam if you want :]

Yes, you may run sa-learn on a folder with real emails, tell it that
those emails are not spam. Bayesian filtering will be even more accurate
that way, but i've never used that feature, i only teach it spam emails.




More information about the Ukfreebsd mailing list