Predictive Modular Neural Networks

This text presents a unified methodology for designing modular neural networks. A family of online algorithms for time series classification, prediction and identification are developed; and a rigorous mathematical analysis of their properties is provided. Case studies involving a number of real-world problems are also presented. Finally, an overview of the modular neural networks literature, including coverage of theoretical and experimental analysis, is provided. The text is a reference work for engineers, computer scientists, and other researchers working in time series analysis, neural networks, control engineering, data mining and other intelligent and decision support areas. The book should also be of interest to researchers in biological and medical informatics.