Data Mining Cookbook: Modeling Data for Marketing, Risk, and Customer Relationship Management (英語) ペーパーバック – 2000/11/3
Kindle 端末は必要ありません。無料 Kindle アプリのいずれかをダウンロードすると、スマートフォン、タブレットPCで Kindle 本をお読みいただけます。
Increase profits and reduce costs by utilizing this collection of models of the most commonly asked data mining questions
In order to find new ways to improve customer sales and support, and as well as manage risk, business managers must be able to mine company databases. This book provides a step-by-step guide to creating and implementing models of the most commonly asked data mining questions. Readers will learn how to prepare data to mine, and develop accurate data mining questions. The author, who has over ten years of data mining experience, also provides actual tested models of specific data mining questions for marketing, sales, customer service and retention, and risk management. A CD-ROM, sold separately, provides these models for reader use.
“…the descriptions are clear, concise, unambiguous…she has clearly succeeded…” (The Institute of Direct Marketing -theidm.com商品の説明をすべて表示する
As the author gives a very brief introduction to data mining, make sure before you even start reading this book that you have a grasp of statistical modelling and data mining in a CRM context, otherwise you will find the material presented in this book too much to take in at once, and worst, you may probably end up being put off building your own data mining applications.
The author clearly has a solid statistical (read SAS) background, making this book a strong contender as one of the best books on data mining around, providing the reader with a number of useful recipes, practical examples and pragmatic data mining approaches which should be studied and understood in detail. Being a cookbook, the author's (or should I say the chef's) particular style may not suite your palate. In other words, you may not like the author's bias towards using logistic regression as the main data mining technique. As a result, you will not learn how to cook exotic dishes using ingredients such as neural networks. However, the choice to use logistic regression as the main statistical techniques pays off, as this allows the reader to start learning to cook robust/reliable meals (models), before cooking with the more exotic ingredients (techniques).
The topics and interventions provided by the well-experienced contributors are in context with the author's material, strengthening the practical context in which data mining applications are presented. On a few occasions, I found that the author does not discuss figures and tabulated outputs in a straightforward way, inevitably affecting the readability of the book. Notwithstanding, the methodology and material presented has a considerable amount of depth and rigour, and the general themes are well structured and maintained throughout.
Many figures and tabulated results are presented in the graphical output provided by the SAS system, which may be less appealing to you if you are not going to be using SAS. Also, many data mining software tools now available have significantly better graphical data presentation capabilities than those presented in this book, inevitably giving it a slightly dated look. Unsurprisingly, being the first version of the cookbook, there are a few typos (and one incorrect figure at the beginning of the first chapter).
In summary, this book is not for the novice, but will be a book that you will want read more than once.
however, some parts of the book were pretty crude. It contains some mistakes. for example, in one chapter the author tries to compare a few repricing scenerios. she compared the account after rate increase with the account before rate increase. and before rate increase, the attrition rate is zero. and it is just not the right way to evaluate a strategy. normally, you would have to compare an account which got a rate raise with the same account as if it didn't receive the increase. and even without the rate increase, the attrition rate down the road can't be zero. normally, you have to use test and control group on this kind of situation. besides, the author made some calculation mistakes in the comparison table. the numbers simply don't add up.
Anyway, overall the book is still a nice one if you can absorb all the nice information in it.