Re: [qmailtoaster] spamassassin bayesian learning (sa-learn)

[Kenneth Swain]

I have been trying to teach toaster spam. I have feed it over 1000 spam and 1500 ham messages. Its spam detection rate is still pretty low. Is there a setting change that might help in this?

On 4/6/06, Jake Vickers <[EMAIL PROTECTED]> wrote:

Riezal Ross wrote:

You may want to have a look at http://www.mail-archive.com/qmailtoaster-list@qmailtoaster.com/msg01304.html

 

Simple explanation is: Create an IMAP folder to put detected spam in it. Feed that folder to sa-learn.

 

On my site (http://v2gnu.com/qmail) I have some info on this as well, as well as a copy of the original script that was sent to me for this purpose. If you need further help, I can send a copy of the modified script that I use.

--
Ken Swain
[EMAIL PROTECTED]
http://kenswain.com
"Spoof them all, let god sort them out."