[Spambayes] SpamBayes as a gatway solution

Itamar Rosenn itamar.rosenn at clinton.com
Wed Aug 27 12:45:51 EDT 2003


Hello,

In Paul Graham's article as well as elsewhere where bayesian spam filtering
is discussed, I've noticed the claim that SpamBayes's technique would not be
effective as a gateway solution.

I'm wondering if anyone has some opinions about whether this is absolutely
set in stone, or whether there might be some good ways to implement
SpamBayes at the gateway level.

I certainly understand the rationale for skepticism; baysian learning is
optimized in an individualized setting, and won't have as many patterns to
follow at the gateway level. But I suspect that at least in my setting,
which is an office of a financial investment company with about 100
employees, there are still quite clear cut patterns of general good mail and
general spam that is received by our network as a whole.

Any thoughts/suggestions on implementing SpamBayes as a gateway solution in
such a setting? (IMAP)

Thanks!

Itamar Rosenn
Clinton Group Inc.




More information about the Spambayes mailing list