Tell me, I have a DF … If I am going to use cross-validation, it is enough for me to break up my DF to the learning and test sample and I do not need to additionally remove the validation set. Right? Or do I understand something correctly?

Answer 1

If you break up your data on the training and test sample and then use cross-validation on this training sample, then the test sample thus obtained is the “validation set” on which you will then additionally be able to check the quality learning quality. It can also be called “deferred sample”.