MLP is a configuration of an ANN that has multiple
layers and is suitable for complicate scenario.
The configuration employs Back Propagation learning
scheme that consists of 2 reciprocal procedures,
i.e., Forward Pass and Backward Pass. The former
passes the data presented at the input layer through
one hidden layer after an-other, while the later adjusts
the connecting weights ac-cording to an error
correction rule, whereby the error signal which is the
difference between the target and actual responses is
fed back into the network in the opposite direction to
the forward links, whose weights are then iteratively
tuned until the actual response reach the specified
target.