A Guide to Neural Computing Applications

CONTENTS
Foreword
Introduction
1.1 Neural computing-today's perspective
1.2 The purpose of this book
1.3 A brief overview
1.4 Acknowledgements
Mathematical background for neural computing
2.1 Introduction
2.2 Why neural networks?
2.3 Brief historical background
2.4 Pattern recognition
2.5 Pattern classification
2.6 The single-layer perceptron
2.7 From the 1960s to today: multi-layer networks
2.8 Multi-layer perceptrons and the error back-propagation
algorithm
2.9 Training a multi-layer perceptron
2.10 Probabilistic interpretation of network outputs
2.11 Unsupervised learning-the motivation
2.12 Cluster analysis
2.13 Clustering algorithms
2.14 Data visualisation-Kohonen's feature map
2.15 From the feature map to classification
2.16 Radial Basis Function networks
2.17 Training an RBF network
2.18 Comparison between RBF networks and MLPs
2.19 Auto-associative neural networks
2.20 Recurrent networks
2.21 Conclusion
3 Managing a neural computing project
Introduction
Neural computing projects are different
The project life cycle
Project planning
Project monitoring and control
Reviewing
Configuration management
Documentation
The deliverable system
4 Identifying applications and assessing their feasibility
Introduction
Identifying neural computing applications
Typical examples of neural computing applications
Preliminary assessment of candidate application
Technical feasibility
Data availability and cost of collection
The business case
Conclusion
Neural computing hardware and software
5.1 Introduction
5.2 Computational requirements
5.3 Platforms for software solutions
5.4 Special-purpose hardware
5.5 Deliverable system
Collecting and preparing data
6.1 Introduction
6.2 Glossary
6.3 Data requirements :
6.4 Data collection and data understanding
7 Design, training and testing of the prototype
Introduction
Overview of design
Pre-processing
Input/output encoding
Selection of neural network type
Selection of neural network architecture
Training and testing the prototype
From prototype to deliverable system
Common problems in training and/or testing the prototype
3 Managing a neural computing project
Introduction
Neural computing projects are different
The project life cycle
Project planning
Project monitoring and control
Reviewing
Configuration management
Documentation
The deliverable system
4 Identifying applications and assessing their feasibility
Introduction
Identifying neural computing applications
Typical examples of neural computing applications
Preliminary assessment of candidate application
Technical feasibility
Data availability and cost of collection
The business case
Conclusion
Neural computing hardware and software
5.1 Introduction
5.2 Computational requirements
5.3 Platforms for software solutions
5.4 Special-purpose hardware
5.5 Deliverable system
Collecting and preparing data
6.1 Introduction
6.2 Glossary
6.3 Data requirements :
6.4 Data collection and data understanding
7 Design, training and testing of the prototype
Introduction
Overview of design
Pre-processing
Input/output encoding
Selection of neural network type
Selection of neural network architecture
Training and testing the prototype
From prototype to deliverable system
Common problems in training and/or testing the prototype
8 The case studies
Overview of the case studies
Benchmark results
Application of data visualisation to the case studies
Application of MLPs to the case studies
Application of RBF networks to the case studies
Conclusions
More advanced topics
9.1 Introduction
9.2 Data visualisation
9.3 Multi-layer perceptrons
9.4 On-line learning
9.5 Introduction to Netlab
Appendix A: The error back-propagation algorithm for weight updates
in an MLP 129
Appendix B: Use of Bayes' theorem to compensate for different prior
probabilities 131
References 133
Index 137

Another Neural Network Books
Download