ABSTRACT
This chapter introduces the mathematical model and structure of artificial neurons. It considers several artificial neural networks that assemble the neurons. The chapter focuses on artificial neurons that simply extract the abstract operation of biological neurons and their mathematical models. It discusses the interpolation mechanism of feedforward neural networks. Artificial neural networks are intended to model the behavior of biological neural networks. The original hope for the development of artificial neural networks is intended to take advantage of parallel processors computing than traditional serial computation. From the graph theory, a complicated network can be always regarded as forming by some simpler networks. Each network has its own “prototype” learning algorithm that achieves the function of storage and recall, learning and mapping, clustering and association, or optimizing a performance index. These prototype learning algorithms have demonstrated the superiority of the design that fits to both network architectures and applications.
