Re: Spam filter reviews?

[Bill Wohler <wohler@newt.com>]

Paul Johnson <baloo@ursine.dyndns.org> writes:

> On Mon, Mar 17, 2003 at 09:29:53PM -0800, Bill Wohler wrote:
>>   Where is this documented? 15 seems a bit high for me; I assume I can
>>   change this in user_prefs.
>
> You'd be amazed how high most people's mothers (and other nontechnical
> family members) tend to get rated by Spamassassin.  15 may actually be
> too low in some extreme cases.

  I've been using spamassassin for some time, and I actually send
  anything with a score of 10 or higher to /dev/null so there is less
  crap to go through when looking for misclassified mail. If's going to
  /dev/null, I think it can safely go into the Bayes db ;-).

  I'd prefer to have it run the Bayes filter for anything over 5 and
  reclassify the odd message that is incorrectly classified as spam
  (only about 1 every few thousand for me, although I suspect that
  bogofilter and spamprobe will have even fewer false positives which
  will make them tempting for me as it will be nice to not even have to
  bother to look for false positives).

  By the way, the default spam score is coincidentally just outside the
  standard deviation from the mean of my corpus of 90 thousand messages.

--
Bill Wohler <wohler@newt.com>  http://www.newt.com/wohler/  GnuPG ID:610BD9AD
Maintainer of comp.mail.mh FAQ and MH-E. Vote Libertarian!
If you're passed on the right, you're in the wrong lane.