On Thu, Nov 25, 2010 at 1:07 AM, ivek gimmick <[email protected]> wrote:
> Hi guys, > > I have a question regarding Bayes classifier. My case is to classify > text as good or bad. > > My question is, do I have to train the classifier with both good and bad > examples? Yes . It will help you to get better results > Or is it enough to train with either of good or bad.? > > It will be something like train a person to identify 'sweet' by giving 'salt' as sample > For eg., I have only bad data. I train the classifier with this data > and generate a model. If I try this on a new set of data, will it classify > something that is not bad? If yes, what would be its label?? > -- ********************************** JAGANADH G http://jaganadhg.freeflux.net/blog
