Random lettered word examples

[David Relson]

On 16 Mar 2004 07:52:20 -0500
Tom Anderson wrote:

> On Tue, 2004-03-16 at 06:47, Tom Allison wrote:
> 
> > X-Bogosity: Unsure, tests=bogofilter, spamicity=0.649109,
> > version=0.17.2 X-Bogosity: Unsure, tests=bogofilter,
> > spamicity=0.503607, version=0.17.2
> 
> > These would have slipped through mine too.  But after one training 
> > they're correct.  I do use a training to exhaustion process on my
> > mailbox.
> 
> If you never get hams in your "unsure" box, or if they are always well
> less than the spam_cutoff, try lowering your spam_cutoff.  I did, and
> my unsure spams were reduced significantly without any false
> positives.
> 
> robx=0.48, robs=0.2, min_dev=0.2, spam_cutoff=0.501, ham_cutoff=0.15
> 
> Tom

My installation is chugging along with:

robs=0.0178, robx=0.549138, min_dev=0.435, ham_cutoff=0.376,
spam_cutoff=0.501

On a typical day I get several hundred spam and several hundred ham and
several unsures.  The unsures are renamed "uh.MMDD.hhmm.txt" or
"us.MMDD.hhmm.txt" for training (as ham and spam, respectively).
Periodically I run "egrep X-Bogosity:.Unsure u?.????.????.txt" to see
what scores I have for my unsure-ham and unsure-spam messages.  I always
see a wide variety of scores -- including spam towards the bottom of the
unsure range and ham towards the top of it.  Much as I'd like to see the
cutoffs tightened to decrease unsures, it's not destined to happen.

David