The mapping m is given by a decision tree able to classify K different classes where the training set used to build this tree is the task . The algorithm used is C4.5 with pruning.