06-04-2014 09:52 PM
Score codes i created by using logistic regression returns with same probablities for every indvidual case when i ran them on test data. It gives perfect scores when ran on training and validation data
What could be the problem
Thank you very much
06-18-2014 04:32 PM
Do you have a constant-only model? That would give you the same predicted value for every observation.