Customer Reviews

Nonparametric Statistics with Applications to Science and Engineering (Wiley Series in Probability and Statistics)

5 star:		(2)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

The authors are both professor at the Georgia Institute of Technology and are accomplished statistical researchers. Vidakovic is an expert probabilist as well, and has also written a probability text on wavelets. He teaches biomedical engineering. Both authors have taught a graduate level engineering course in nonparametric statistics and they both have done some research in nonparametric methods. This text is very modern as it includes bootstrap methods, Bayesian nonparametric methods and wavelets with an eye toward engineering applications. The first five chapters are simply a review of basic concepts in probability and statistics, then in Chapter 6 goodness of fit methods are covered. Chapters 6 - 10 cover the standard topics. Interesting features are the introduction of pictures of famous statisticians who have contributed methods with their names associated with them. Mann and Whitney, Kruskal and Wallace, Fisher and Friedman are among the ones that belong in this group.

Chapter 11 covers density estimation and Chapter 12 covers robust regression, isotonic regression and generalized linear models. The remaining chapters cover curve fitting, wavelets and the bootstrap with the engineering models and applications and are very valuable modern techniques with engineering applications. Statistical learning methods are introduced and sometimes are important to engineers and engineering applications. This is an excellent text and could be a useful reference for engineers and statisticians.

I will be reviewing this book in the future for JASA and will be more detailed in my coverage there.

After reviewing several texts for Nonparametric Statistics and dozens in "Statistics for Engineers", this is the *only* book that actually has me excited. [Honorable mention goes to Terrell: Mathematical Statistics: A Unified Introduction.] All the other books -- both traditional nonparametrics and engineering stats -- are anachronistic rehashes of classic texts or classic approaches. YAWN!

These classic texts are excellent background material, and what Kvam and Vidakovic have realized is that you can do the material from them in a fraction of a semester and get on with relevant topics. Yes, you need to know distributions. Yes, you need to know rank methods. Now let's get on with the other 80% of what an engineering analyst should know in the twenty-first century! (With the notable exception of Information Theory -- so sprinkle in some Shannon if you want.)

The authors have the sense to realize that engineers use MATLAB. You can use Minitab if you want; but engineers use MATLAB. You can use R if you want (as a statistician, I would); but engineers use MATLAB. You can use SAS if you want; but engineers use MATLAB. In case you haven't noticed, engineers use MATLAB!

For my taste, I would like a slightly more "mathematical statistics" text (engineers can still write integral signs, yes?) but I can take this text and add some math here and there a whole lot easier than I can take any of the standard score of texts and make them *relevant*. Great job!