On Mon, 2008-07-14 at 15:48 +0200, Stefan Jakobs wrote:
> On Friday 11 July 2008 17:29, andys wrote:
> >   for a mail server running email for multiple domains what is the
> > typical/recommended way to collect emails which arent detected as spam to
> > be processed by sa-learn? Users are downloading mail via POP3, so once a
> > users sees a mail and decides that it is in fact spam its already been
> > removed from the mail server. If the user forwards the mail to a special
> > mailbox for processing then the mail is obviously now different from the
> > original spam, the user is the sender etc. Will sa-learn still work using
> > this method? and if not what else can I implement that would work?
> 
> This is what I do:
> Forwarding the unrecognised message to an account which will process the 
> message through sal-wrapper.pl. You will find further informations here: 
> https://po2.uni-stuttgart.de/~rusjako/sal-wrapper

Forwarding alters the message, you will not get reliable results.

You can, of course, use auto-learn and let SA take care of it.

If you want your users to classify, the best way is to use IMAP instead
of POP, and provide server-side training folders that sa-learn can see.
If IMAP is not an option then this obviously won't work.

If procmail is in use as the LDA, you could set up a rule to clone to a
local ham folder to do scheduled training. You could get creative with
rules and have it collect a randomly-chosen subset of the ham traffic,
or only train where the score is low and the message is not already
BAYES_00 or the score is high and the message is not already BAYES_99.
However, this would be cloning users' mail (even if only temporarily),
and you should obtain their consent before doing this.

-- 
 John Hardin KA7OHZ                    http://www.impsec.org/~jhardin/
 [EMAIL PROTECTED]    FALaholic #11174     pgpk -a [EMAIL PROTECTED]
 key: 0xB8732E79 -- 2D8C 34F4 6411 F507 136C  AF76 D822 E6E6 B873 2E79
-----------------------------------------------------------------------
  Usually Microsoft doesn't develop products, we buy products.
                          -- Arno Edelmann, Microsoft product manager
-----------------------------------------------------------------------
 2 days until the 63rd anniversary of the dawn of the Atomic Age

Reply via email to