An ideal inductive machine learning algorithm produces a model best approximating an underlying target function by using reasonable computational cost. This requires the resultant model to be consistent with the training ...