This is pretty confusing. What has happened is that you have encoded a single categorical variable that has four states as four numerical variables. Unfortunately, Mahout has gotten the message that you are using four categorical variables each with two states.
Can you say a bit about how you ran the command? Also, take a look at chapters after 13 for more examples that don't just use numerical attributes. On Tue, Dec 13, 2011 at 3:30 PM, magicalo <[email protected]> wrote: > Hello, > > I am trying to understand the output of TrainLogic. Mahout in Action only > has an > example of running the classifier on numeric predictor variables. However, > my > model uses categorical predictor variables only (a,b,c,d) and each can > only hold > a value of 0 or 1 only. The output I get is this: > > myslassifier ~ -63.081*a=0 + -15.640*a=1 + > -47.268*b=0 + -31.457*b=1 + > -47.269*c=0 + -23.541*c=1 + > -23.635*d=0 + -15.729*d=1 + .... + > -23.635*Intercept Term > a=0 -63.08073 > a=1 -15.63973 > b=0 -47.26775 > b=1 -31.45702 > c=0 -47.26934 > c=1 -23.54059 > d=0 -23.63467 > d=1 -15.72897 > > How should I interpret this? Why do I have 2 weights for the same predictor > variable (e.g.: a=0 and a=1) and why are all of the weights negative? I > have not > been able to find any documentation on this. Thanks! > > Thanks a lot! > >
