Christopher M. Bishop
This book provides a solid statistical foundation for neural networks from a pattern recognition perspective. The focus is on the types of neural nets that are most widely used in practical applications, such as the multi-layer perceptron and radial basis function networks. Rather than trying to cover many different types of neural networks, Bishop thoroughly covers topics such as density estimation, error functions, parameter optimization algorithms, data pre-processing, and Bayesian methods. All topics are organized well and all mathematical foundations are explained before being applied to neural networks. The text is suitable for a graduate or advanced undergraduate level course on neural networks or for practitioners interested in applying neural networks to real-world problems. The reader is assumed to have the level of math knowledge necessary for an undergraduate science degree.
1 Sheer pleasure.
If you want a very good, intermediate introduction to pattern classification this book must be on your bookshelf. It even does a very nice job explaining the EM algorithm in a few pages! Basic calculus is all you need to understand the book. A must read.
2 It makes a difficult topic easy to understand
The theories of NN and PR are quite difficult to understand. But this book makes them much easier. The author can explain the concepts without using too much formula. If other authors could follow his step then the life is much easier!
3 Recomended book to read
This is a recommended book to read for people who would like to read about statistics and maths. People with few knowledge about these sciences will find it a bit difficult to read.
4 An excellent book
When I came across this book, I had already read several on the subject, including Beale & Jackson (lightweight) and Haykin (middleweight)
For a reader unafraid of basic statistics and linear algebra, this is an excellent beginning book. For the math wary, I would say read a math-lite conceptual book first. This was a text book in my master's program, and I heard from students with a weak math background that they found it extremely challenging.
Bishop rightly emphasizes the statistical foundations of feedforward networks. This is a large subject in and of itself, and he covers it well. It provides an extremely solid foundation.
Neural dynamics via recurrence, Hopfield Nets, and many other topics outside or on the edges of feedforward networks are not covered.
I find many NN books are poorly written, imprecise, and have little content. This is one of the best books I have read on the subject.
5 good book but pity is that it do not have a disk accompy it
Strongly suggest the author include matlab scripts for his example and problem.
6 Believe me -- there is no better book for beginners
This is definitely the NN bible for beginners. I used it first in 1996 just after it came out and I still use it for reference. Reading some of the other reviews I saw that some people think there is too much math in the book -- that is not true -- the well explained math in the book is necessary to make the topic extremely clear.
Now 6 years later it would be nice to have a second, extended edition covering other successful NN related areas like recurrent NNs, PPCA, ICA, etc., also maybe some online adaptation techniques using Bishop's gift of being able to explain in simple words & math.
7 Disappointing from my point of view
I was looking forward to a detailed insight into neural networks in this book. Instead, almost every page is plastered up with sigma notation (which gets incredibly tedious after the first 5 chapters). The author relies on formulae too much to impart his obviously vast expertise onto the reader. I am an above average computer programmer, and I understand most of the concepts in this book, since I have come across them in my other research projects. But Bishop does very little to explain some of the important key concepts in this book. For example, why so much mileage can be gotten out of integrals and sequences in this field is mystifying. Practical knowledge if NN's are not only assumed but taken completely for granted!
I imagine that a great deal of knowledge can be gleaned from this compendious tome, but it is very hard reading.
Good use of graphics though. I just wish this book was twice as long, so he could have fitted in some more background info. Then and only then will this book be worth its £26 asking price in my opinion.
I just hope it comes in useful one day!
8 Very good work
This book is the best treatment of the subject. To really understand the content, it's necessary prior knowledge of probability theory, but not in depth. It is well illustrated and, more important, the topics are explained in manner logic and sweep. This work don't contains everything, but it's cool because it's readable and sufficient rigorous.
9 Very formal and well presented
Although this book is not for beginners, you can use it as a startup text as long as you can understand the math behind it. The contents are beautifully presented and with the expected detail and formalism of such a great book. As a software developer I also use other books that are more algorithm-centered, but this is the one I look for when I want to read a formal exposure.
10 NOT FOR BEGINNERS
This book is not for beginners. It is heavy into the mathematical side of neural networks. I bought this book hoping to be able to take away the overall picture, a more conceptual overview, but my ignorance in math prevented that. In one sentence:
Way over my head.
I'll give it 5 stars because the people who could understand it seem to think the world of it :)
11 fine technical exposition
I found the clarity of the math and technical aspects of pattern exposition to be extremely high. The more math, in particular statistics, one has the better, but still does an excellent job in explaining some of the basic concepts for those who have not had sufficient exposure to them.
Certainly fundamental and I would consider a valid university text.
12 An excellent introduction to pattern recognition
Do not be put off by the title: this book is more about pattern recognition than neural networks. Of course it covers neural networks, but the central aim of the book is to investigate statistical approaches to the problem of pattern recognition.
An excellent companion to "Duda & Hart".
As other reviewers have said: you will need a reasonable maths or stats background to get the most out of this book.
13 Grows on You
This book came out at about the same time as Ripley's, which has almost the same title, but in reverse. At the time, I liked Ripley's better, because it covered more things that were totally new to me. Then a friend said he had chosen Bishop for a course he was teaching, and I went back and reconsidered the two books. I soon found that my friend was right: Bishop's book is better laid out for a course in that it starts at the beginning (well, not quite the beginning--you do need to be fairly sophisticated mathematically) and works up, while Ripley's is more a collection of insights all at the same level; confusing to learn from. Bishop is able to cover both theoretical and practical aspects well. There certainly are topics that aren't covered, but the ones that are there fit together nicely, are accurate and up to date, and are easy to understand. It has migrated from my bookcase to my desk, where it now stays, and I reach for it often.
To the reviewer who said "I was looking forward to a detailed insight into neural networks in this book. Instead, almost every page is plastered up with sigma notation", that's like saying about a book on music theory "Instead, almost every page is palstered with black-and-white ovals (some with sticks on the edge)." Or to the reviewer who complains this book is limited to the mathematical side of neural nets, that's like complaining about a cookbook on beef being limited to the carnivore side. If you want a non-technical overview, you can get that elsewhere, but if you want understanding of the techniques, you have to understand the math. Otherwise, there's no beef.
14 Excellent mathematical reference of Neural Networks
A good book if you are looking for learning mathematical teory of Neural Networks or set a parameters of comercial application. Not recommended for beginners
15 Extraordinarily well written and comprehensive
Rarely do I encounter a book of such technical quality that also is a pleasure to read. Bishop moves through sometimes difficult topics in a clear, well-motivated style that is appropriate as both an introduction and a desktop reference on neural nets. Definitely on the "A list."
Bishop chose to not include discussions on a number of topics that might have diluted his focus on pattern recognition (for example, Hebbian learning and neural net approaches to principal components analysis). I think that these choices greatly strengthened the integrity of his presentation.
I would love to see an updated edition with a discussion of recent results in statistical learning theory, kernel methods and support vector machines.
16 Excellent technical reference and tutorial
I'd like to agree with previous reviewers. Note that you will need a good mathematical background (especially in statistics) to understand the content. However, the book is completely thorough in developing all the key concepts and really tries to give you insight into the meaning behind the equations. It's style is that of an undergraduate level textbook, but a very well written one. To use neural nets effectively, I think you need to have at least one book like this.
17 A Thorough and Rigorous Introduction
This is a terrific book if you want to understand why neural nets work, and how to make them work. As advertised, it really goes into practical issues like preprocessing and generalization, which are easy to do halfheartedly, but are complex issues if you really want to get the best results. If I had to have only one book on neural nets, this would be it, no contest.
18 Just the right blend of intuition and mathematical rigor.
Bishop cuts through the hype surrounding neural networks, and
shows how they relate to standard techniques
in statistical pattern recognition. He concentrates on feedforward
and radial basis function networks, which are the ones used most
widely in practice. This book is about as mathematical as
Hertz, Krogh and Palmer ("An Introduction to the Theory of Neural
Computation", 1991), but is probably easier to read, and is
certainly of more use to the practitioner. A real gem!
