Just wanted to verify a line of thinking I was planning to include in the
overview doc. We have to give some indication to verifiers on how to use the
information.
I think that it is important to point out that you are not going to get good
results from naïve learning schemes and make sure that there is sufficient
coverage there to defend against the objection.
-----Original Message-----
From: Jim Fenton [mailto:fenton(_at_)cisco(_dot_)com]
Sent: Wednesday, August 23, 2006 3:12 PM
To: Hallam-Baker, Phillip
Cc: J.D. Falk; ietf-dkim(_at_)mipassoc(_dot_)org
Subject: Re: [ietf-dkim] Bayesian filters are the pits
While an interesting topic, this seems only peripherally
related to DKIM and I wonder whether a different venue, such
as the ASRG list, would be more appropriate.
-Jim
Hallam-Baker, Phillip wrote:
I think that I am working towards a rather broader critique
of the way that SpamBayes &ct. are applied.
Naïve Bayesian learning schemes are intrinsically
vulnerable to counter-programming. They work on a small scale
only because there is not a sufficient value to counter-programming.
...
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html