There were a lot of negative comments for this book on Amazon so I sort of held off the purchase of this book. However, last week I ordered the book and quickly realised it is very useful. The book may lack detailed theory but excellent references are scattered throughout if you really want to follow it up. And the book is easy reading and well written with most statistical concepts defined before they are used.

The focus of this book primarily is to explain how to work on statistics using Matlab and it provides a taste of various areas with adequate explanations and code to get started. One advantage of this book is they do not define their own notation but use the notation which is currently in vogue in academia.

If you are starting out in Matlab, are not a statistician and do not have previous experience with other packages (like Splus or R) you should definetly think about getting a copy. If you are a Stats Guru you can just read the toolbox documentation. However note that these authors provide their additional stats toolbox FREE (which is also well written) on the website which contains most of Matlab statistical functions so you could save yourself some money on the Stats toolbox.

I think there would be a real interest in a book on "computational statistics" and related topics that showed details of analyses and algorithms using Matlab. This book is expensive and extremely disappointing.The explanations are sparse and very weak and the m.files are usually small add-ons to functions from the Stats Toolbox.

I think in any book on this topic there have to be detailed explanations of how methods work and what their limitations are.Otherwise the reader can find themselves in a lot of trouble very quickly. There is insufficient detail either for a student coming to the topics for the first time or for someone actually wanting to analyse data.

Other books that people might want to have a look at:
1)Statistical Pattern Recognition 2nd edition . Andrew Webb.This is not oriented to any particular language.Good introduction.
2)Netlab. Ian Nabney (this has excellent Matlab functions for neural networks)
3)Modern applied statistics with S 4th edition, Venables and Ripley. This uses a different language (but which will be relatively easy for Matlab users to learn), but learning S or R (free!) makes a huge number of tools available.
4)The recent data mining book by Hand et al. This offers clear and cogent explanations.It is good for someone who does not want overly mathematical descriptions.

I haven't looked properly at the recent Hastie,Friedman and Tibishirani book yet, but you can find reviews on the Amazon page for the book.

My major complain was that the authors, in general, did not present algorithms clearly. Limit selections of algorithms did not help either. As a result, you cannot use this book as a reference because it just does not contain enough material. You cannot learn much about computational statistics with this book becuase the statistic methods and algorithms are not adequately presented. you cannot even write codes for your own statistic analysis with the MATLAB examples shown in the book unless you have the Statistics Toolbox.
The only persons that might be benefit from this book are those who don't want to read the Statistics Toolbox manual on line. Given that the Mathworks no longer ship printed manuals, this book may be used a companion of the Statistics Toolbox.

I ordered this book assuming to get something useful, but i got the impression while reading the book that it is like a collection of notes from other books, wrapped with some matlab code. More worrying to me was that the mathematics makes a sloppy impression. For me that means I cannot grab the book to lookup something and use the code without having to be concerned on the validity. The bottomline being I will not use it for applications and the book is a waste of money.

As an independent student of probability and statistics this was a great find. You get a great overview of the useful algorithms of computational statistics. The chapter on Exploratory Data Analysis with its use of multidimensional graphing was very enlightening. It's wonderful that each topic is accompanied by source code (free on-line) that lets you see exactly how it's done. It's easy to tweak the code and explore your own data as you go along. You get just enough theory to understand the algorithm and lots of good common sense and rules-of-thumb on how to best apply it. Finally, the extensive bibliography and chapter-by-chapter annotations will point you straight to the best source for more in-depth study.