Browse Books

Reviewer: Yannis Manolopoulos

This second edition of this textbook has been expanded to cover topics that emerged from recent research. It discusses the concepts of data mining, but also includes a complete reference to the topics of data warehousing and online analytical processing (OLAP). Nowadays, excessive amounts of data are being generated in domains as diverse as market research, functional genomics, and pharmaceuticals. Analyzing this data, with the aim of answering crucial questions and helping make informed decisions, is the challenge addressed by the concepts discussed in this book. Data mining (also known as knowledge discovery) is the process of analyzing data collections from different perspectives and summarizing them into useful information. Data warehousing is a process of centralized data management and retrieval. All of these technologies are becoming indispensable parts of business intelligence. The book starts by explaining what data mining is. Next, the authors address the data itself in terms of its quality, usability, and organization for efficient access. They then explain what a data warehouse is, and, in particular, how it differs from a database. The next chapter (a new addition since the first edition) examines data cube generation and data generalization issues. The central chapters, 5 through 8, address various data mining tasks, such as association rules discovery; classification and prediction; cluster analysis; and mining stream, time-series, and sequence data (this latter task is also new material). In the final two chapters, 9 and 10, useful extensions are presented, examining complex data such as graphs, social networks, text, multimedia, spatial data, and World Wide Web data. Recent work in these emerging research fields is summarized. A clear introduction to data mining is presented in the book. It is nicely laid out as a textbook, with an orderly summary, problem set, and bibliography at the end of each chapter. The book consistently uses data from a fictitious organization to illustrate most concepts. This gives a strong sense of cohesion to what can actually be very different techniques. One key aspect of the book is its question-and-answer (Socratic) format. The good thing about this format is the clean way that it introduces a new topic, something readers will appreciate. The bibliography helps the reader decide which references will give the best description of each of the chapter's topics. The book is also very usable, as it provides pseudocode for many techniques and prepares the reader for more advanced discussions. This is one of its merits-instead of going into great depth, it offers readers an easy explanation of the big picture of each topic and points them in the right direction (in the related bibliography) for further understanding. The book can be used for both students and practitioners. We have been using it for several years as an introductory-level book for a master's of science course on data mining and data warehousing. It requires a basic background in mathematics, a fact that makes it appealing to students with various backgrounds in computer science. Little normal statistical analysis is used, though some facility with statistics will help. However, reasonably good programming skills are required. For practitioners, this is one of the few books on data mining that contains material of technical merit, in the sense that it explains how data mining algorithms actually work. This is a good starting point for any implementation. Moreover, the second edition provides some coverage of data mining tools. A great number of data mining books either contain a shallow description of the involved techniques or are nothing more than improperly titled books on machine learning. This textbook, one of the few on data mining algorithms and techniques, has become a standard work. For these reasons, I recommend it very highly. Online Computing Reviews Service