[Spambayes] Is Equal Ham & Spam really the best?

[David Abrahams]

on Sun Jul 29 2007, skip-AT-pobox.com wrote:

>     Dave> Even training only on mistakes and unsures, I have had a steadily
>     Dave> increasing ratio for months.  I almost never see a misclassified
>     Dave> ham and only very rarely a ham about which the system is unsure.
>     Dave> It's unsure about spam every day.
>
> What are your ham and spam thresholds set to?  

The default, whatever that is, for classification.  And I've been
using

  -o Categorization:ham_cutoff:0.02  \
  -o Categorization:spam_cutoff:0.98 \

when I invoke tte.py

> Have you tried lowering your spam threshold a bit?

No, I haven't tried that.  I suppose I run the risk of a few false
positives if I go too far with that?

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com

The Astoria Seminar ==> http://www.astoriaseminar.com