[Spambayes] Bayesian drawbacks?

Gerrit Holl gerrit at nl.linux.org
Mon Sep 29 12:57:38 EDT 2003


Tim Peters wrote:
> > Is it possible that a non-Bayesian approach would much better suit my
> > needs,
> 
> You didn't explain what your needs are, Gerrit.  If your primary criterion
> is "doesn't need any training", then sure, a filter that needs training
> isn't for you.  I suggest you just try it, or try to make clearer what your
> needs are.

I am convinced to do so by the (several) replies I got. I guess I was probably
unjustly distrusting something of which I did not understand how
it could possibly work. Maybe I should offer my apologies for this
skepticism <g>. I was also a little bit scared off by the many
warnings about how not to use it (evenly distributed ham/spam), by the warning
that spam changes so that I would keep feeding it mail and by stories
of a non-bayesian spamfilter. I have now started trying it out: I will see
whether it suits by needs, it probably will, judging by other answers
to my question.

Thanks for all the answers!

Gerrit.

-- 
165. If a man give to one of his sons whom he prefers a field, garden,
and house, and a deed therefor: if later the father die, and the brothers
divide the estate, then they shall first give him the present of his
father, and he shall accept it; and the rest of the paternal property
shall they divide.
        -- 1780 BC, Hammurabi, Code of Law
--
Asperger Syndroom - een persoonlijke benadering:
	http://people.nl.linux.org/~gerrit/
Het zijn tijden om je zelf met politiek te bemoeien:
	http://www.sp.nl/



More information about the Spambayes mailing list