NEURAL NETWORK PC TOOLS

Reviewed 5/13/1992

Neural Network PC Tools, by Russell C. Eberhart and Roy W. Dobbins

NEURAL NETWORK PC TOOLS: A Practical Guide
Russell C. Eberhart (editor)
Roy W. Dobbins (editor)
San Diego: Academic Press, November 1990

Rating:

5.0

High

ISBN-13 978-0-12-228640-7
ISBN 0-12-228640-5 440pp. HC/GSI $44.95

As a novice in the world of neural-network computing, I had long searched for a book which would give me three things: 1) a solid, yet not overly theoretical introduction to the field; 2) a comparison of the performance of the various types or models of neural networks; and 3) some examples of code that I could try out on my PC at home. This book provides all three.

The book is applications-oriented throughout, and contains many well-constructed diagrams to illuminate the techniques it describes. Chapter 1 gives an overview of the history of the field, beginning about 100 years ago with William James, and evaluates the most influential books and papers. Chapter 2 describes implementations of the back-propagation model and the self-organizing model, and Chapter 3 covers the specifications of a problem to be solved and offers advice on avoiding the pitfalls that may occur.

The remainder of the book moves away from this system-level discussion into actual code and on to the hardware. These chapters, beginning with Chapter 4, are: Software Tools; Development Environments; Hardware Implementations; Performance Metrics; Network Analysis; and Expert Networks. The chapter on hardware (Chapter 8)1 is devoted mainly to the Inmos T800 Transputer. Following this, five case studies are presented:

A glossary of terms and seven pages of densely-typed references are provided. Finally come appendices containing source code listings in C (and one in Turbo Pascal) and a list of additional sources such as organizations, conferences and journals. The source code is also available on disk from one of the editors for $20 (or $26 outside the U.S.)

This book is an excellent resource for anyone working with neural networks. Beginners like me will find it invaluable in getting up to speed on the subject. Even old hands, I think, will still discover useful information in it.

1 By the list of chapter titles, this should be 6. When I can get to my copy of the book, I will check it.
Valid CSS! Valid HTML 4.01 Strict To contact Chris Winter, send email to this address.
Copyright © 2003-2014 Christopher P. Winter. All rights reserved.
This page was last modified on 7 June 2014.