Neural Network Principles (Second Edition)

Author: Hai Jin
Publisher:
Publish Date: 2004-01-01
Features: Neural networks are one of the most active branches in the research, development, and application of computational intelligence and machine learning. This book is a standard textbook on neural networks, providing a comprehensive and systematic introduction to the basic models, methods, and techniques of neural networks, with in-depth research on both the basic models and main learning theories of neural networks. It offers particularly detailed and systematic analysis of the derivation of learning theories and learning algorithms, and provides a comprehensive and integrated introduction to the new development trends and main research directions in neural networks. The theoretical and practical applications are closely integrated, laying a solid foundation for the practical application of neural networks, making it an extremely readable textbook. Neural networks are one of the most active branches in the research of computational intelligence and machine learning. This book comprehensively and systematically introduces the basic concepts, system theory, and practical applications of neural networks. The book consists of four parts: Introduction, Supervised Learning, Unsupervised Learning, and Neural Network Dynamic Models. The Introduction part covers neuron models, neural network structures, and the basic concepts and theories of machine learning. Supervised Learning discusses perceptron learning rules, supervised Hebbian learning, Widrow-Hoff learning algorithms, backpropagation algorithms and their variants, RBF networks, regularized networks, support vector machines, and committee machines. Unsupervised Learning includes principal component analysis, competitive learning in self-organizing feature mapping models, information theory in unsupervised learning, random learning machines rooted in statistical mechanics, and finally, reinforcement learning related to dynamic programming. Neural Network Dynamic Models study dynamic systems composed of short-term memory and hierarchical feedforward networks, the stability and associative memory of feedback nonlinear dynamic systems, and another class of nonlinear dynamic-driven recursive neural network systems. This book 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. The book includes numerous examples and exercises, along with 13 computer experiments based on MATLAB software. This book is suitable as a textbook for graduate students or senior undergraduate students, and can also serve as a reference book for scientists and engineers who wish to delve deeper into neural networks.