On Mon, Jun 21, 2004 at 06:51:14AM -0700, Jeffrey Goldberg wrote:
|
| Well, yes and no. With the exception of aol, earthlink and amazon, the
| bulk of the "pass"es that I see are from spammers. So I fully expect
| that my statistical filtering (Bayes with autolearn through
| spamassassin) is learning to treat passes as weakly
| spammish.
Machine learning algorithms need to key off the combination
of authentication-result and sender-domain. Doing one
without the other is like eating a sandwich without any
bread.
http://spf.pobox.com/slides/unified%20spf/0335.html