I've read in multiple places that calibration on model results shouldn't be done on the training set (the set that the model is build on), but rather, on a set that the model have not seen. I failed to understand the reasoning.

I tried Googling this but I've seen a lot of conclusions not explanations. Could someone hint me on the reasons behind it?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.