Re: classifier questions

"Eric S. Johansson" <[email protected]> Mon, 11 May 2009 23:02:29 -0400
Newsgroups gmane.mail.spam.crm114
Message-ID <[email protected]>
[email protected] wrote:
>
> 
> 2) the Hyper, OSB, OSBF and Markovian score distributions all
> appear to be roughly gaussian to the eye, centered about zero
> (though OSB and Markovian have a few way-out-there outliers)
> 
> Is a gaussian distribution expected for these classifiers?

As they say in the classics, "your mileage may vary.".  I've had horrible luck
and my OSBF score distributions are what a very kind person would call
"bimodal".  I get a huge blob of scores below zero and a smaller, rounder blob
of scores above zero.
> 
> 3) Winnow's score distribution is "bell-ish" in shape, but with
> a peak well to the positive of zero.
> 
> Is this expected?

Yes.  In fact, the osbf classifier also has a DC offset of about three.  I need
to spend some time on integrating graph generation in my antispam system so I
can track how the scores are going south.

------------------------------------------------------------------------------
The NEW KODAK i700 Series Scanners deliver under ANY circumstances! Your
production scanning environment may not be a perfect world - but thanks to
Kodak, there's a perfect scanner to get the job done! With the NEW KODAK i700
Series Scanner you'll get full speed at 300 dpi even with all image 
processing features enabled. http://p.sf.net/sfu/kodak-com