Hofmann and Klinkenberg have produced a fine collection of essays on data mining and analytic models, presented in several cross-disciplinary cases. This book describes data mining and case applications using Rapidminer models and analytic techniques (http://www.Rapidminer.com).

Rapidminer is a system for the design and documentation of an overall data mining process. The system offers a comprehensive set of operators and structures that can be used to express the control flow of the process using drag-and-drop tools. The book focuses on the fine details of Rapidminer, from data preparation to model development to evaluating and visualizing the results; further support is freely offered at the website.

The book represents the work of more than 30 contributors. Managing the writing styles of so many contributors is a challenging task, and the editors are to be commended for their effort. The material flows well, is very readable, and easily transitions from chapter to chapter and section to section. The book is divided into ten sections, each focusing on a different disciplinary area and a different analytic and mining model. Each section includes one or more cases.

Section 1 introduces Rapidminer and data mining in general. Section 2 discusses basic classification, using cases in credit approval, teaching assistant selection, and nursery school selection or rejection. The classification models consider k-nearest neighbor classification and naïve Bayesian classification. Section 3 explores cases in marketing, cross-selling, and recommender systems for higher education programs. Section 4 focuses on medical and educational cases, with a focus on clustering algorithms. Section 5 discusses the mining of text rather than numerical variable data, such as in spam detection, language identification, and customer feedback. Rapidminer includes a text processing extension.

Section 6 covers feature selection and classification in astroparticle physics and carpal tunnel syndrome in medicine. A feature selection extension is available in Rapidminer. Section 7 presents models and mining results for datasets in molecular structure and the modeling of property-activity relationships in biochemistry and medicine. Section 8 focuses on image mining, including feature extraction, segmentation, and classification. Section 9 examines anomaly detection, instance selection, and the construction of prototypes. Finally, Section 10 models meta-learning and automated learner selection.

In each section, the authors introduce a mining activity, the related model, and the analytic techniques used. The datasets are described, the Rapidminer requirements are enumerated, and the analysis is summarized. The reader can refer to the companion web site for downloadable code and datasets for each case (http://www.rapidminerbook.com). If you would like more information about the book, this is the place to look.

Data mining requires a basic knowledge of clustering and classification algorithms, linear models, principal component analysis, and factor analysis, among other grouping and discrimination techniques. A background in such topics would be helpful for reading the book. Similarly, some technical savvy will be useful. Tutorials, manuals, and other support material for Rapidminer are available at http://rapid-i.com.

This is a good book. If you are interested in some very interesting data mining cases, or if you would like to learn Rapidminer, it will not disappoint. The bibliographic references are lengthy and the indices are well done.

More reviews about this item: Amazon