ABSTRACT
Artificial neural networks (ANN) are designed to mimic the architecture of the human brain in order to create artificial intelligence like human intelligence. This chapter introduces two types of ANNs for classification and prediction: perceptron and multilayer feedforward ANN. It explores the processing units and how these units can be used to construct various types of ANN architectures. The chapter describes multilayer feedforward ANNs with the back-propagation learning algorithm and provides a list of software packages that support ANNs. Processing units of ANNs can be used to construct various types of ANN architectures. The number of inputs and the number of outputs in an ANN depend on the function that the ANN is set to capture. The chapter discusses a graphical method and a learning method to determine connection weights for a perceptron, which is a one-layer feedforward ANN with a sign or hard limit transfer function.
