Task: Train a Backpropagation network to classify a given dataset.
Write a Backpropagation network whose units take 3 inputs (X, Y, Z).
Because a Backpropagation network can only output values less than 1 you may choose to divide the range (0, 1) into several areas, each being associated with a different class.
Alternatively, you may wish to have as many networks as classes. So, if you have 3 classes and are 5 training a network for finding class 2, you could have 0.75 as target if class is 2 and 0.25 for target for the other classes.