Several advances in techniques for obtaining nonexplicit information that is present in database systems are described in this book. It discusses a good number of problems and techniques from the point of view of what the authors call market basket databases.

The book does not present a unified theory of data mining. Instead, it is organized as a set of independent chapters that are almost self-contained papers on one topic, including a separate reference list for each chapter. The chapters have some internal loose connection in themes and notation.

The starting point of the book consists of the “classical” papers of data mining [1,2], which are briefly reviewed when presenting the first advanced topic in the book: conditional patterns. Another common point in all of the chapters lies in considering the pre-existence of a set of frequent itemsets, as is common in the data-mining literature.

The usual topics of knowledge discovery in databases are revisited with an expanded view. So, for example, going beyond the usual association rules, the topics of conditional patterns and Boolean expressions are covered. The topic of measuring associations between data items is also expanded with several new measures. Then the main topics of the book are tackled: mining several databases, time-stamped data, and joining different data sources.

Throughout the book, a similar pattern is used to present each chapter. It starts with a presentation of another problem of data mining, usually related to a previously mentioned problem, some theoretical discussion, an algorithm, and experimental results using a few standard benchmarks.

The content is interesting for readers with specific interests in data mining, but one would expect a little more care in the presentation of the material. In the initial chapters, there can be found tables with wrong entries, references pointing to the wrong sections, and lemmas/theorems with confusing statements. However, all of these problems are of a lesser nature, and all of them could have been eliminated with a detailed reading of the material by the authors.

With respect to the book’s content, the algorithms presented are also sometimes unrealistic, as they contain iterations over an exponential number of items, without the expected discussion on the limits of their applicability. Studying the algorithms, one is often led to believe that the subsequent experiments were developed over a small set of items, so as not to reach its intractable limit.