http://bugzilla.spamassassin.org/show_bug.cgi?id=2910





------- Additional Comments From [EMAIL PROTECTED]  2004-01-22 20:11 -------
Subject: RE:  Fast SpamAssassin score learning tool.



> 
> perceptron.   I took Henry's advice and tweaked the parameters a 
> little to see
> what effect that would have.

When you have a chance, would you please re-upload the latest/greatest
perceptron scores (the ones you like the best from testing)?

Also, I was wondering if you could upload the "ranges" file that you
were using with the GA? I was curious about looking at divergences from
the mean of those ranges.

And ... is there a write-up, in the archives or somewhere regarding the
thinking behind the TCR and lambda calculations?

> PS: I dropped the mystery extra column.  I haven't a clue what 
> that was for. ;)
> 

Good. Had me worried. <g>

Excellent work, Henry, Sidney, JM, all. I think that with a more efficient
scoring calculation, it will leave more room/time for experimentation and
tuning, and at some point will make it possible for those of us with our
own custom rules, and our own corpi to be able to tune the scores to our
set up without having to reserve time on a Cray.





------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.

Reply via email to