Re: [OFFTOPIC] spam scoring (was: *****SPAM***** Extracting a KeySym from an action routine)

On Tue, Nov 26, 2002 at 01:45:30PM -0500, Branden Robinson wrote:

>Hmm, so we have 3 pieces of evidence that it isn't spam, four that it
>is (all of which are from blacklists), the blacklist scores were enough
>to get the message scored as spam, and yet it wasn't spam.
>
>Are we sure we want to be investing this much trust in RBLs, at least
>for this mailing list?  Or is there a mountain of spam that hasn't made
>it to the list thanks solely to the RBL rules that makes this false
>positive worth the trouble?

I've been collecting data for the last month or so using two spam
filtering tools (spamassassin and a home-grown one that we already use
for our private lists).  After I've analysed that data, I'm planning to
use one or both of those filters to tag or drop mail sent to our public
lists, and remove the subscriber-only posting restriction.  As can be
seen from the spamassassin-tagged messages that got through, we're not
yet filtering based on that information.

All I can really say so far without having analysed the data is that
the number of false positives has been relatively small compared to the
number of valid positives.  I need to assess now many valid positives
were attributable to the RBL matching before coming to any conclusions
about that.

David
-- 
David Dawes
Release Engineer/Architect                      The XFree86 Project
www.XFree86.org/~dawes