New new script to train bogofilter
Boris 'pi' Piwinger
3.14 at logic.univie.ac.at
Fri Jul 4 16:24:45 CEST 2003
David Relson wrote:
>>Actually, I don't know why randomtrain needs more messages.
>>But I don't understand that script.
>
> My experiments confirm that randomtrain does more training than does
> bogotrain.pl. It would be interesting to know why. bogotrain.pl does an
> orderly test ham, then spam, then ham, then spam, ...
The reason is that I need some order. If I'd do all spam
first, then all ham, I'd expect way more errors. This way I
learn "at the same time" from spam and ham.
> randomtrain is, well, random.
Also random if it takes spam or ham next?
> It seems that either train-on-error is highly dependent on
> message order
You mean the number needed? What is the exit condition for
randomtrain? Mine (-f) is that there are no errors left.
pi
More information about the Bogofilter
mailing list