Question

I’m training Weka’s logistic regression classifier and I’m trying to figure out what is going on under the hood. I know that I can use the classifier to look at the confidence distribution per instance using the logistic.distributionForInstance method but is there a way that I can look at the feature weights learned by the classifier?

thanks

Answer 1

Yes you can get the weights as a 2D array of doubles with the coefficients method. The first index is for the attributes and the second index is for the classes.