Graham's method seemed better
Greg Louis
glouis at dynamicro.on.ca
Sat Nov 23 13:39:14 CET 2002
On 20021123 (Sat) at 1217:31 +0100, Boris 'pi' Piwinger wrote:
> Greg Louis <glouis at dynamicro.on.ca> wrote:
>
> >We can conclude that it is unlikely to be true that Graham's
> >calculation method misses fewer spams than Robinson's, or than the
> >variant of Robinson's that is based on Fisher's method of combining
> >probabilities.
>
> I certainly did not understand the complete analysis. Maybe
> I should elaborate on how I had my experience:
<snip>
> evidence of an effect. But as you know I used bogofilter out
> of the box and did not tweak anything.
Rem acu tetigisti. You can't expect to do that successfully, any more
than you can run a car without using the steering wheel. Especially
with geometric-mean Robinson, adjusting the spam_cutoff level to the
actual mail population is crucial. Robinson-Fisher is said to be less
touchy, but I'm using 0.952 in my implementation of that because it
works better than 0.95 in my production system.
> PS: Is any of the native speakers still shocked by the use
> of a plural form of the word mail?
Dunno what it would mean. There's no such thing as "a mail" even
though there's "an email." So how can there be mails?
--
| G r e g L o u i s | gpg public key: |
| http://www.bgl.nu/~glouis | finger greg at bgl.nu |
More information about the Bogofilter
mailing list