To compare classifiers it would be handy to have a machine-readable version
of the confusion matrices, instead of just a text printout.

On Sun, Sep 11, 2011 at 10:36 AM, Robin Anil <[email protected]> wrote:

> Its much simpler we don't need any of the tokenizing portions in old
> TestClassifier. Simply multiply document vector with class vector. And sort
> the score pick the highest for NB or lowest for cnb
> On Sep 11, 2011 10:12 PM, "Grant Ingersoll" <[email protected]> wrote:
> >
> > On Sep 11, 2011, at 11:50 AM, Robin Anil wrote:
> >>
> >> Unfortunately I havent made a wiki entry yet. I will put one it soon.
> The
> >> naivebayes package only has the model building code. Once the model is
> >> built, the class vectors need to be loaded and the scores need to be
> >> computed for each new document (dot product). This part is not yet
> there.
> >
> > You just mean simply load the model and do the scoring, right? Can we
> reuse what's in TestClassifier or any of that stuff?
>



-- 
Lance Norskog
[email protected]

Reply via email to