Hi, I am trying to do URL categorization. I have generated NB model and deployed in web service. my NB model able to classify on the fly url categorization. For evaluation I have taken 80-20 (training-testing). As of now accuracy is 62.39%.
I would like to increase the accuracy up to 70 %. Could you help me on this. Thanks, Venkat On Fri, Jun 6, 2014 at 11:17 PM, Ted Dunning <[email protected]> wrote: > All, > > Translation for the rest of us: 1 lakh = 100,000 > So this is 2.8 million URL's. > > Venkat, > > Can you say what you were trying to do? > > > > > On Fri, Jun 6, 2014 at 5:20 AM, venkata ramana < > [email protected]> > wrote: > > > Hi Mahout, > > > > I have crawled the 28 lacks of url s. I were able to get 62.39 % of > > Accuracy through Naive Bayes. I am trying feature Dictionary to tweak my > > model. > > Could anyone please let me know some of the techniques to tweak the NB > > Model. > > > > > > > > Thanks, > > Venkat > > >
