min_dev vs spam_cutoff [was: spam cutoff less than neutral? ]
Boris 'pi' Piwinger
3.14 at logic.univie.ac.at
Tue Feb 24 07:50:31 EST 2004
David Relson wrote:
> The last step of the Robinson-Fisher algorithm, which is what bogofilter
> users, is a reverse chi-square test. This test answers the question
> "for the number of tokens scored, and the spamicity seen, what is the
> likelihood that this message is spam?"
Worse than that. The test only says something about the
failure of a hypothesis. This test only works in one
direction. So we need another test for the opposite
direction. Those two results are then combined into one
number which is *scaled* to fit the 0,1-interval.
More information about the Bogofilter