explaining Bogofilter simply

Bill McClain wmcclain at salamander.com
Sun Jan 25 18:37:04 CET 2004


On Sun, 25 Jan 2004 17:06:32 +0000
Richard Kimber <rkimber at ntlworld.com> wrote:

> The message he quotes ("Highland alberich rampart discovery barnet
> clothesman walpole boot brainwash ...") would only be classified as
> real mail by a decently trained Bayesian system if one's
> correspondence normally contained this combination of words.

You might give your spamicity score of his example sentence, as in "On
my system, the bayesian filter gives this message a score of 0.96, which
is deep into spam territory and close to a perfect 1.0, denoting pure
and unmistakable spam"

And: "Contrary to rumour, neither random nor garbage words in a message
help it escape from a trained Bayesian filter."

-Bill
-- 
Sattre Press                            Curiosities of the Sky
http://sattre-press.com/                    by Garrett Serviss
info at sattre-press.com        http://sattre-press.com/csky.html




More information about the Bogofilter mailing list