In supervised learning, we have an existing Training Set of "right" answers to train the algorithm on.
A supervised learning problem is either a:
