countdown boutiques-francophones Learn more vpcflyout Pets All-New Kindle Music Deals Store sports Tools Registry

Customer Reviews

3.5 out of 5 stars
3.5 out of 5 stars
5 star
4 star
3 star
2 star
1 star
Format: Hardcover|Change
Price:$105.00+ Free shipping with Amazon Prime

Your rating(Clear)Rate this item
Share your thoughts with other customers

There was a problem filtering reviews right now. Please try again later.

TOP 100 REVIEWERon February 23, 2013
I read this book as the text in a four-week online class on cluster analysis. I learned a great deal and do not regret purchasing this book. It has several strengths and some weaknesses as an introduction to this statistical technique.

There is a good introduction to the unsupervised learning problem of classifying objects into meaningful groups with no basis for validating these classifications. The authors' decision to focus on graphical methods early in the text is a good one and lays an intuitive foundation for their more technical presentation later in the book. The discussion of similarity measures at the core of cluster analysis is a good overview and prepares readers for more advanced discussions elsewhere.

The book closes with the highly useful and practical chapter "Some final comments and guidelines." It lists and describes nine steps in a typical cluster analysis and refers readers back to sections of the book which inform the decisions at each step. It's coverage of methods for testing cluster quality and the likelihood of no structure in a dataset is also accessible and of practical value. Readers might consider looking through this material before reading the previous chapters to help organize the information more meaningfully.

The middle chapters are worth reading, but suffer from a few problems. In general, these chapters are better at describing the boundaries of current research in clustering techniques than they are in describing typical applications. There are too many research results and not enough examples. The examples that are included are described too briefly, making it difficult to follow how the analysis was carried out. Better integration of citations in the body of the text would be a key improvement. As would inclusion of sample exercises with worked-out solutions in an appendix.

Recognizing the difficulty of making a statistics text accessible to readers using a variety of software packages, I still believe this was not done well in this book. See Iain Pardoe's Applied Regression Modeling for one example of how to do this very well. I will hope for improvements in a later edition of this book.

This book has challenges as a text, but was worth the price and the time spent with it. Still, I will be on the lookout for a better alternative.
0Comment|Was this review helpful to you?YesNoReport abuse
on April 15, 2003
Here is an excellent introduction to cluster analysis. The concepts are explained in clear language, with many illustrative examples. It is possibly the best of the introductory level books. I give it 4 stars because of a few misprints, and a few places where some essential information or detail has been omitted that can lead to misunderstanding.
0Comment|Was this review helpful to you?YesNoReport abuse