This decade has seen an explosive growth in computational speed and memory and a rapid enrichment in our understanding of artificial neural networks. These two factors provide systems engineers and statisticians with the ability to build models of physical, economic, and information-based time series and signals. This book provides a thorough and coherent introduction to the mathematical properties of feedforward neural networks and to the intensive methodology which has enabled their highly successful application to complex problems.