Neural network design

Author: Dai Kui
Publisher:
Publish Date: 2002-09-01
Features: This book introduces the basic structure and learning rules of neural networks, with a focus on the mathematical analysis, training methods, and applications of neural networks in engineering practice such as pattern recognition, signal processing, and control systems. The material is organized in a clear and consistent manner to facilitate easy reading and use. For each discussed topic, numerous examples are provided to illustrate the concepts. Since this is a book about neural network design, the selection of topics follows two principles: first, to prioritize practical neural network structures, learning rules, and training methods; and second, to ensure the book's completeness, allowing readers to progress smoothly from one chapter to the next. To this end, introductory materials and chapters on the fundamentals of applied mathematics are included before specific topics. In summary, among the topics we have chosen, some are extremely important in the practical application of neural networks, while others are highly useful for explaining how neural networks operate. This book primarily covers the fundamental concepts of neural networks, introducing practical network models, learning rules, and training methods. The book is divided into 19 chapters, covering topics such as neuron models and network structures, perceptron learning rules, supervised Hebbian learning, Widrow-Hoff learning algorithms, backpropagation algorithms and their variations, associative learning, competitive networks, Grossberg networks, adaptive resonance theory, and Hopfield networks. It emphasizes the discussion of mathematical analysis methods and performance optimization, highlighting the application of neural networks in practical engineering problems such as pattern recognition, signal processing, and control systems. Additionally, the book includes numerous examples and exercises, along with "Neural Network Design Demonstrations" programs based on the MATLAB software package. This book can serve as a textbook for senior undergraduate or first-year graduate students taking neural network courses, as well as a reference for researchers and professionals engaged in related fields.