My solution is to have everyone make a folder called SpamLearn and HamLearn.
Then daily (weekly) run :
sa-learn --spam /home/vpopmail/domains/*/*/Maildir/.SpamLearn/cur/
sa-learn --spam /home/vpopmail/domains/*/*/Maildir/.SpamLearn/new/
sa-learn --ham /home/vpopmail/domains/*/*/Maildir/.HamLearn/new/
sa-learn --ham /home/vpopmail/domains/*/*/Maildir/.HamLearn/cur/
Users who have these folders, get their spam (and ham) learned.
After that it would be nice to delete those messages, but i haven't
looked into it how to do it so quota doesn't break of something like that.
Jake Vickers wrote:
[EMAIL PROTECTED] wrote:
Hi guys,
I've been searching but can't find the best way with QMT to create a
folder then maybe run the bayes learning technic on it...
[sa-learn --showdots --mbox --spam spam-file]
I was thinking of have a [EMAIL PROTECTED] account for clients to
forward
mail into if it is spam etc. Than run the bayes learn against that.
I'm new to all of this.
Can someone please point me in the right direction? Links? Howto?
There's some info on the wiki about this - most clients like Outlook
and Thunderbird will rewrite the headers when you forward the message,
so you're just teaching your Bayes that messages from you are spam.
You can use the spam box and run a check on that (there's a script in
QTPlus to do this) or you can just copy the messages at the CLI to a
separate folder.
---------------------------------------------------------------------
QmailToaster hosted by: VR Hosted <http://www.vr.org>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
