Data Mining Techniques in CRM: Inside Customer Segmentation Review

Average Reviews:

(More customer reviews)This book is an excellent guide for the practitioner, a data mining specialist who needs to do something with his/her customer information. It covers basic machine learning algorithms like kmeans, decision trees and self organization maps, pca, etc. I recommend this book as a starting point for people who don't have any experience with data mining and statistics and they want to do something with their data. It is a big plus for the book that it gives hints about how to choose the options for algorithms, since this is very critical and hard to find. We as a company use this book to understand the problems that the business world faces. I think the book has the right size and sufficient number of examples that are very well explained.
Although I understand the need of picking a tool to express the examples I think sticking to SPSS so tightly gives a bias to the book of what is feasible and what is not. I think some examples with open source like R would help the book to be less biased. But again I don't think it is a major issue, since the majority of the audience is professionals that will buy SPSS.
My major objection though, has to do with the distinction of machine learning and statistical learning that authors make. As it is stated in page 61 decision trees and neural networks are machine learning methods and not statistical ones, while pca is a statistical one and not a machine learning one. In reality all of them are statistical methods, machine learning and statistical learning are the same thing in the literature. A more sensible taxonomy between statistical methods is parametric/non-parametric. It is true that in general parametric methods can be faster versus non-parametric, but the statement in the book that they are more accurate is not in general valid. Nonparametric methods are slower but they are by far the most accurate. In reality though, nonparametric methods have been accelerated recently and they can actually be as fast as parametric ones.
I definitely recommend this book and I think the authors have done very good job, it filled a gap between science and business.
Nikolaos Vasiloglou
CTO Analytics 1305

Click Here to see more reviews about: Data Mining Techniques in CRM: Inside Customer Segmentation

Get 21% OFF

Buy cheap Data Mining Techniques in CRM: Inside Customer Segmentation now.