In statistical machine learning, the standard measure of accuracy for models is the prediction error, i.e. the expected loss on future examples. When the data distribution is unknown, it cannot be computed but several resampling methods, such as K-fold cross-validation can be used to obtain an...