Artificial Neural Networks

Abstract

Artificial neural networks (ANN) are a form of computing for predicting a set of outputs from a set of input patterns (a set of variables). This machine learning method is motivated by the functioning of biological neural networks that consist of neurons and links (nerves). The neurons serve as the processing units and links serve as the connections between neurons. A commonly used form of ANN is a multiple‐layer feedforward network in which neurons are organized into layers: one input layer, zero, one or two hidden layer(s), and one output layer. The links connect the neurons from the different layers (input to hidden and hidden to output). Most ANNs just have one hidden layer. The neurons in the input layer receive input patterns and send those inputs through the links to the neurons on the hidden layer, which then process the inputs and send the results to neurons in the output layer for processing and presentation. The application of ANN consists of two general steps: network training and network application. Network training uses samples, each of which contains an input pattern and the expected output pattern for that sample, to develop associations between a set of input patterns and a set of output patterns. The network application step uses the trained network to predict an output pattern for a given input pattern whose corresponding output pattern is unknown.