The principle of neural networks derives from the human brain in which neurons are fired to process information gathered from a stimulus.
Neural networks take an array of input values, normalised between zero and one, and pass them through a series of layers containing nodes.
Each node also holds a value between zero and one, and the relationship between nodes of one layer and the next is determined by a matrix of weights and biases.
A single node in a layer is dependent on every node in the previous layer multiplied by their respective weight and offset by a bias (37).