Introduction to Neural Networks for Java, 2nd Edition [Paperback]

Jeff Heaton (Author)

Book Description

Publication Date: October 1, 2008

Introduction to Neural Networks with Java, Second Edition, introduces the Java programmer to the world of Neural Networks and Artificial Intelligence. Neural network architectures, such as the feedforward, Hopfield, and self-organizing map architectures are discussed. Training techniques, such as backpropagation, genetic algorithms and simulated annealing are also introduced. Practical examples are given for each neural network. Examples include the traveling salesman problem, handwriting recognition, financial prediction, game strategy, mathematical functions, and Internet bots. All Java source code is available online for easy downloading.

Editorial Reviews

About the Author

Jeff Heaton is an author, college instructor, and consultant. Jeff lives in Chesterfield (St. Louis), Missouri. The author of seven books and over two dozen journal and magazine articles, Jeff specializes in Internet, socket-level/spidering and Artificial Intelligence programming. Jeff has worked with companies such as MasterCard, Anheuser-Busch and Boeing. A Sun Certified Java Programmer and a Senior Member of the IEEE. Jeff holds a Masters Degree in Information Management from Washington University in St. Louis.

Product Details

• Paperback: 440 pages
• Publisher: Heaton Research, Inc.; 2 edition (October 1, 2008)
• Language: English
• ISBN-10: 1604390085
• ISBN-13: 978-1604390087
• Product Dimensions: 9 x 7.5 x 1 inches
• Shipping Weight: 1.6 pounds
• Average Customer Review: 3.5 out of 5 stars  See all reviews (4 customer reviews)
• Amazon Best Sellers Rank: #839,343 in Books (See Top 100 in Books)

More About the Author

Jeff Heaton is an author, consultant, artificial intelligence (AI) researcher and former college instructor. Heaton has penned more than a dozen books on topics including AI, virtual worlds, spiders and bots. Heaton leads the Encog project, an open source initiative to provide an advanced neural network and bot framework for Java and C#. A Sun Certified Java Programmer and a Senior Member of the IEEE, he holds a Masters Degree in Information Management from Washington University in St. Louis. Heaton lives in St. Louis, Missouri.

Customer Reviews

Most Helpful Customer Reviews

7 of 7 people found the following review helpful:

5.0 out of 5 stars Good book for practical programming, November 30, 2009

By 

Matthew Skoda (Pearl Harbor, HI United States) - See all my reviews
(REAL NAME)   

Amazon Verified Purchase

This review is from: Introduction to Neural Networks for Java, 2nd Edition (Paperback)

I've read several author's works on neural networks, and while valuable, they almost always limit the subject to theory.

This book limits it's scope to practice, which is fine by me.... If you want theory, feel free to look it up on wikipedia or one of the billion books on AI that only cover theory.

This book assumes you've read a little theory and jumps straight into practice: in it, the author walks you from hands-on from creating the basic neural nodes to creating and training simple decision nets, to building applications for predicting stocks moves and playing backgammon.

I've ready about neural nets, but was unsure on how to apply them in practical applications: this book clarified their design and usage. However, I will warn you, it's not an easy read, and requires you to have the code loaded on your computer nearby... this book is about practice, and the author pumps a lot of information out.

14 of 18 people found the following review helpful:

2.0 out of 5 stars Lack of focus; doesn't deliver, October 13, 2009

By 

Eric Hackman (Bay Area, CA, USA) - See all my reviews
(REAL NAME)   

This review is from: Introduction to Neural Networks for Java, 2nd Edition (Paperback)

Jeff Heaton undertakes the admirable task of writing a book that provides some background in neural network theory overlaid with a layer of practical Java coding how-to. Unfortunately, the book delivers on neither of these intentions particularly well. I have only a very basic background in neural networking (see "Neural Networks: A Comprehensive Foundation - 2nd edition" by Haykin) and was expecting Mr. Heaton to provide some theoretical explanation of why various network architectures are relevant to particular types of problems, how training algorithms differ from each other and why this is important...actually any bit of theory at all to shed some light on what the code was supposed to do. Instead, there's a brief chapter on matrix math, which in my opinion was not especially clear, and an occasional ball and stick diagram and almost no theoretical explanation of where any of it comes from or why it's important. I learned next to nothing about neural networking from this book, and what I may have learned is probably available on wikipedia in the space of a few paragraphs.

I could resign myself to the lack of neural networking explanation if the book instead presented a robust discussion of Java design as applied to neural networking architectures and algorithms. But, alas, this is not to be found either. The Java code is presented with no insight into the author's design decisions and therefore offers little in the way of educational material. Unless you are truly a Java novice, the code in the book will seem obvious and underwhelming. While it's apparent that basic neural networks can be constructed with relatively simple code, the author's failure to provide any explanation of code design or to frame the code within the context of a larger neural networking library perhaps results in the Java coding how-to portion of the book failing to deliver as well.

In short, I read the first 6 chapters of this book and decided not to waste any further time with it. If you want to understand neural networks, you won't find that here. If you want to learn to write Java code to build neural networks, you won't find that here either. You'll find code that the author has already written that you can use, but there won't be much educational value in it. The book truly is more of a user's manual or technical documentation for the author's neural networking Java classes and not much more. Perhaps that is useful if you want something simple you can drop into a project and run with. My suggestion to those who wish to learn and understand how to build neural networks in Java is to learn a little about the networks themselves then hack out some Java code yourself. You'll understand what the code means and be in a much better position to extend that code. And you'll definitely learn something along the way, which, unfortunately, I did not while reading this book.

4 of 4 people found the following review helpful:

3.0 out of 5 stars software + insights, August 18, 2010

By 

Charles E. Knadler Jr. "math biologist" (Salt Lake City, UT USA) - See all my reviews
(REAL NAME)   

Amazon Verified Purchase

This review is from: Introduction to Neural Networks for Java, 2nd Edition (Paperback)

The usefulness of this book depends on your purpose. If your primary purpose is to acquire a running start solving AI problems using the author's software this is a useful book. If you want to study neural networks or to develop java programming skills, this is probably the wrong book for you. The author gives good introductions to genetic algorithms and simulated annealing as well as their use in training neural networks. There are also valuable insights for the careful reader; e.g., the approach used for the tic tac toe game (chapter 6) has very limited usefulness, because it took "nearly 20 hours to train" the neural network. This raises the question of how long it would take for a more realistic problem. (The author provides the weights, allowing the reader to experiment with the neural network, without the pain of performing the training.) I found the chapters on genetic algorithms and simulated annealing to be the most useful.

The author'a website provides downloads of all his software with liberal licensing.

For self-study, "Elements of Artificial Neural Networks" by Mehrotra et al or an older book, "Fundamentals of Neural Networks" by Fausett, would be an excellent companion to Heaton, if you would like a combination of accessible theory with a working software suite that can be readily modified to solve real problems.