../sT-fold cross-validation (Stone 1974) repeats k times for a sample set randomly divided into k disjoint subsets, each time leaving one set out for testing and the others for training. Thus, we may call this technique “leave some out”