Radical lexers

[michael at optusnet.com.au]

"Boris 'pi' Piwinger" <3.14 at logic.univie.ac.at> writes:
> michael at optusnet.com.au wrote:
> 
> >This isn't super suprising. You're testing with a small corpus,
> 
> Not that small, that you could not see large differences in
> the errors.
> 
> >on a very easy data set. 
> 
> Why do you think so?

It looks like you're taking all the ham and spam from a single
mailbox.

This changes the classification task from 'find that spam'
to just "find things that don't looks like my normal
email".

I think. :)

Michael.