Testing fisher

[Greg Louis]

On 20030129 (Wed) at 1415:50 +0100, Boris 'pi' Piwinger wrote:
> Greg Louis wrote:
> 
> > I note that your totals have grown; one improbable but not impossible
> > artefact could be that you've introduced a bunch of hard spams in the
> > most recent batch.
> 
> This is unlikely, since I just added the normal flow of mail.
Yeah, I didn't think it very likely.
 
> > Trouble is, I can't think of a mechanism by which raising robx, in the
> > absence of any other parameter changes, could lower resulting
> > spamicities.  Any chance you could repeat the experiment, just for
> > 0.025, 0.020 and 0.015, with the exact same training db and message
> > corpora, with both robx values?  We could be onto something big here.
> 
> I can with the same training set. The bogofilter database is
> constantly updating, though. If you also want to avoid this,
> I could make a copy of those and freeze it this way.

That would be the way to go.  A prime rule of experimentation,
especially when an unexpected finding has been obtained, is to
eliminate all avoidable extra factors.  (Same idea as when your
computer breaks you pull out all the cards and put them back one by one
till the defect appears.)  Running with absolutely nothing different
but the robx value will force us to interpret results in terms of robx
instead of finding excuses ;)

Regards............
-- 
| G r e g  L o u i s          | gpg public key:      |
|   http://www.bgl.nu/~glouis |   finger greg at bgl.nu |
| Help free our mailboxes. Include                   |
|        http://wecanstopspam.org in your signature. |