From: "Lars Oeschey" <[EMAIL PROTECTED]>
> I got my SA finally running, though there's still a wealth of spam
> getting through. I don't know if I should first investigate on the rules
> for SA or start the bayes filter.
> If I run the bayes filter, is there an easy way, to train it from a imap
> client? For example, I create a Mailfolder "Ham" and a folder "Spam"
> (already got that anyway where Spam gets in from SA), ans sa-learn scans
> those folders from a cron-job? I hav courier imap with Maildir format,
> is that possible?
>
> Lars
Um, that is what I am doing on an RH9 install that serves two people
on Windows machines using Outlook Express.
Main delivery is via a POP3 server. The IMAP server is used as a tool
for moving, mostly, unidentified spam into a spam mbox file that is
used for automated training each evening. I even put together a little
futility that automatically strips down the spam mailbox and appends
the stripped off data to two files, one used for training and the
other for long term storage of our respective corpus bundles. I also
setup respective ham and oldham folders for periodic training with
known ham.
I can put together a "bundle" for you as guidance for setting up
such a trick for yourself, if you wish. (The futility is open source,
"If it works I wrote it. if it doesn't I don't know who wrote it."
So far it works.)
{^_^}
