After building a classifier with modern tools of machine learning we typically have a black box at hand that is able to predict well for unseen data. Thus, we get an answer to the question what is ...
No takes yet. Share an insight, caveat, or question.
BaehrensDavid et al. (2010) studied this question.