I have a classification model which I initially built in Azure ML studio and then created a similar model in python (its similar because the algorithms in python VS Azure are a bit different so they are not exactly the same). What I am observing is that the average scored probabilities for both classes is very different between the two models. The azure results are more definitive which means the ones which were categorized as class 1 have probabilities a lot closer to 1 and the reverse for class zero. Python one is interesting because the probabilities are more normally spread between 0 and 1. both results are pretty similar in terms of my desired metric so I can use both of them but I wonder why there is this huge difference between these two models. I appreciate any hint.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.