Brian Azzopardi wrote:
I agree with you on the effectiveness of statistical filters for
personal use.
However, I wouldn't be so quick to dismiss them for corporate use. My
observations are that most corporate email is internal and that only a
relatively small portion of employees need to receive work-related mail
from external sources - in particular sales, marketing and other outward
facing staff. So for small to medium companies it works great.
Large multinations and ISPs are a different matter altogether...
Yeah ! The real problem isn't personal or not, but how much the
population is homogeneous. Statistical filters may work very well with a
very big population, as long as the trafic is homogeneous.
This is an optimisation problem - for a big population you have many
local optimum values for the filter and global optimum values. The more
all these values are near, the more your bayesian filter will be
efficient on a big population. Unless you use one set of parameters per
user...
Brian Azzopardi
-----Original Message-----
From: asrg-bounces(_at_)ietf(_dot_)org
[mailto:asrg-bounces(_at_)ietf(_dot_)org] On Behalf Of
Jose Marcio Martins da Cruz
Sent: Tuesday, September 21, 2004 11:37 AM
To: laird(_at_)lbreyer(_dot_)com
Cc: ASRG
Subject: Re: [Asrg] Filtering spam by detecting 'anti-Bayesian'
elements?
...
You're completely right on all this affirmations. What I really wanted
to say is that statistical filters are very, very good for personnal
use, but not for corporate structures. And most of the time I hear about
bayesian/statistical filters, this particular point is left...
--
---------------------------------------------------------------
Jose Marcio MARTINS DA CRUZ Tel. :(33) 01.40.51.93.41
Ecole des Mines de Paris http://j-chkmail.ensmp.fr
60, bd Saint Michel http://www.ensmp.fr/~martins
75272 - PARIS CEDEX 06
mailto:Jose-Marcio(_dot_)Martins(_at_)ensmp(_dot_)fr
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg