Gives an introduction to the mathematical properties of feedforward neural networks and the computationally intensive methodology that has enabled their highly successful application to complex problems of pattern classification, forecasting, regression, and nonlinear systems modeling. This work is also of interest to those in computer science.