<< 1 >>
Rating: 
Summary: Neural Networks for Intelligent Signal Processing
Review: ***** Stanley McGibney, Consultant Mechanical Engineer I found this book to be significantly different in its treatment of neural networks for signal processing and pattern recognition. It deals very ably not only with essential theory but also with basic practical issues, often missing from other books on the subject, that significantly enhance understanding and application. Zaknich has included a nice guide and design approach to successful application of neural networks, which is supplemented by frequent tips and a variety of worked application examples. The book is much more than a good introduction to neural networks. It also includes a class of neural networks that Zaknich has developed and worked on over a decade that he refers to as common bandwidth spherical basis function neural networks. This is based on a generalization called the Modified Probabilistic Neural Network (MPNN) that encompasses Donald Specht's Probabilistic and General Regression Neural Networks. He has continued to develop the MPNN in a number of very practical directions that allows it to used for a wide range of engineering problems. He seems to favour applications related to underwater acoustic signal processing but the methods and approaches that he offers are suited to many other non-linear problems found in engineering and other disciplines. The book includes a very interesting discussion on intelligent signal processing. Zaknich talks about what he calls hyperspace signal processing in the context of the MPNN and other classical filtering structures that gives an interesting view of some of the basic issues involved. He suggests at least one possible generic approach to non-linear signal processing based on Vapnik's Support Vector Machine that has a structural similarity to the MPNN. This book is a gem that shines in its clarity beyond many other books on neural networks that I have struggled with in an attempt to understand the subject well enough to apply it.
Rating:
Summary: Neural Networks for Engineers
Review: Zaknich's book provided me with the necessary theory and detail to enable me to develop a neural network application as part of my phd. I thoroughly recommend this text for anyone wishing to develop and use neural networks, particularly for engineering applications. I found that while most books focussed on either the theory of neural networks or selected applications, Zaknich's text provided a comprehensive coverage of both theory and application providing a sound basis for understanding and applying neural networks. Topics ranged from the applicability of neural networks through to data collection, data conditioning and the final implementation.
Rating:
Summary: Neural Networks for Engineers
Review: Zaknich's book provided me with the necessary theory and detail to enable me to develop a neural network application as part of my phd. I thoroughly recommend this text for anyone wishing to develop and use neural networks, particularly for engineering applications. I found that while most books focussed on either the theory of neural networks or selected applications, Zaknich's text provided a comprehensive coverage of both theory and application providing a sound basis for understanding and applying neural networks. Topics ranged from the applicability of neural networks through to data collection, data conditioning and the final implementation.
<< 1 >>
| 
