We will adopt the most common NN type, the multilayer perceptron, where
neurons are grouped into layers and connected by feedforward links (Fig. 2).
Supervised learning is achieved by an iterative adjustment of the network connection
weights, called the training procedure, in order to minimize an error
function. For regression tasks, this NN architecture is often based on one hidden
layer of H hidden nodes with a logistic activation and one output node with a
linear function