Mastering Data Mining: The Art and Science of Customer Relationship Management

Although aged the book remains a precious guide for business and CRM people. It argues that data mining is a discipline that must be mastered by concentrating in how to approach analytics than how to use tools. Specifically it stresses the importance of: - the iterative nature of data mining activities (and project life cycle) - the active involvement of business people - the business objectives and needs - the preparation and split of the model set (mining view) - the evaluation of produced models and patterns - the business interpretation of data mining results - the power of data exploration (by example) "Mastering Data Mining" is a much more concrete and comprehensive book than "Data Mining Techniques, 2nd edition".

While doing a graduate elective on Decision Making Technologies, I realized that data visualization and representation is crucial for data exploratory and validation of data mining analysis. To get some ideas on how the various data visualization and workflow techniques are applied and integrated into the GUI of commercial softwares, survey the various chapters of this book.

This book is an excellent book. The authors explain the various techniques, and show real world examples of their use. Most importantly, they explain the underlying goals of the various techniques, and what to watch out for when using them. I was most relieved to read that I am not alone in having limited success with association rules! Although some of the particular examples were not the type of examples I deal with, the reasons they were chosen make perfect sense. Data mining owes much of its popularity to people attempting to find churners, etc. But there are plenty of examples covered, and with each one some new insight is revealed. Especially useful to me were the explanations of what it is one sees in the decision trees, lift curves, etc. Also, seeing various problems solved with several of the popular tools (MineSet, Enterprise Miner, etc.) was very helpful. There are many examples from various industries, and you learn something new about those industries too! (If you like the Sesame Street videos of how cans, tires, etc. are made even more than your kids do, you'll love this book for the examples alone.)It is clear from this book that the authors not only know what they are talking about, they can actually break it down for a newbie like me. I have also had the pleasure of being in one of Mr. Berry's MineSet classes, and he demonstrated the same depth of knowledge and ability to convey it to others in that class as well.This book is not an algorithm book, but it touches on them. It is not necessarily a tour of data mining tools, but does do this to some degree. It is probably most useful for anyone who wants to know "What is this 'data mining', and how can it help me?" with real world examples to make things clear. If the reader starts out thinking that data mining is just tossing a bunch of data into a tool and getting concrete results back, the confusion will not remain after reading this book. Finally, this book is VERY easy reading. Do yourself (or your boss) a favor and buy this book!

Many books have been written on the algorithms used for data mining (e.g., machine learning, statistics). This is not yet another one.This book is geared at people who want to derive insight and take action in a business setting. It is now well known that the algorithmic step is only a small part of the iterative knowledge discovery process, yet few books enlighten the users with the issues involved.This book has a small section on the algorithms, but concentrates on the often-overlooked PROCESS of data mining (sometimes called knowledge discovery) and the problems associated with this process in practice. Michael and Gordon are practitioners who have used multiple data mining tools and techniques. They know the problems and describe them well, sharing their real-life experiences through actual case studies. For example, people rarely appreciate the main problem with association algorithms: the number of uninteresting rules they generate. Now I can show them pages 426-428.The few things that I didn't like were the use of non-standard terminology in a few cases. For example, directed instead of supervised; prediction instead of regression. While the common terms aren't great, they're standard now. The book also has few references. Someone readers will want to read more details about specific areas and will not find needed references.Overall, it's a well written book, easy to read, with nice analogies to the world of photography.

Michael and Gordon are two of the best data mining consultants in the business and in this book they provide a thorough description of how to turn the technology into action. This book expands on their fantastic first book (Data Mining Techniques) with a wide variety of case studies and insight into what it takes to use data mining to impact real business problems. Highly recommended.