Re: classifier questions

Jason White wrote:

> This is interesting. Even after fixing my MUA configuration to pipe messages
> to mailreaver with CRM114 headers intact, and starting with newly initialized
> CSS files, after some training, OSBF starts classifying all messages as
> "good", with quite high scores.
> 
> HyperSpace (which I'm now trying) seems to be better, but after several days
> of training, most messages are still marked as "unsure", although the actual
> scores appear to be tending in the right direction.
> 
> Any suggestions as to how I can get substantially higher accuracy would be
> welcome. I am training all errors and unsures as I encounter them.

How about a joint effort of building a  test set and tools to process and score?
 I keep all my messages in an sqlite database and I'm sure we can find some
common format/ground through which we can show these errors etc.

one word of warning, I do everything in Python because I have nuked hands and
must use speech recognition to write code (and damn near everything else)

------------------------------------------------------------------------------
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