Customer Reviews

A User's Guide to Principal Components

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

 

 

This book is an excellent choice for helping understand data compression and noise reduction of large datasets. It is extremely beneficial, especially when dealing with hyperspectral datasets, to understand the techniques involving the transformation of multiple bands into principal components. The book is well organized according to the general method(s) by which PCA works. From the compression of information content in a multiple number of bands, to other uses of principle components analysis, this is definitely an excellent reference for anyone who works with hyperspectral data.

I find Jackson's book to be well-written and in a style that is almost conversational. He gives sound advice for stepwise evaluation of characteristic roots and residual analysis in Chapter 2. I have really only skimmed the surface with this book, but so far I like what I have read and am satisfied with the purchase.

I find Principal Component Analysis (PCA) a perfectly usable technique that has a place in a statistical toolbox. It is an unfortunate fact that in many applications areas, PCA has become the de-facto Multivatiate Analysis Technique, in some cases even becoming synonymous for that term. In an ideal world, a book like Jackson's would simply not be necessary. If more sophisticated analysis was required to solve a problem, any number of techniques far more powerful than PCA can be brought to bear. However, there is a user community that wants to augment PCA with multiple layers of secondary analysis and interpretation, and this book is for them.

Having stated my dislike for the need for this book, I concede that it meets that need quite well. It is written in an approachable manner, presents simple data sets, and is a little bit less math intensive than some of the more general machine learning texts.

This book is geared toward engineering types, not for people who want to use PCA for stock trading, securities covariance forecasting, etc.

I've received this book maybe 3 or 4weeks ago.
But I found that there is one problem.
The last several pages of that book are torn and folded.
I decided not to claim anything...
but I want for you to be more careful of your things(products).
Anyway, thank you for the delivery.
Sincerely yours.