Association rule mining. Models and algorithms (Q1600090)

Association rule mining: models and algorithms deals with recent research in the area of association rule mining. Following an introductory chapter on data mining in general, Chapter 2 presents basic concepts of and algorithms for association rule mining. The following five chapters deal with negative association rules, causality and causal rules, association rules in very large and in small databases. The book closes with conclusions and directions for future research. As stated in the preface, the book is basically intended for anyone who is interested in association rule mining. Despite this broad target audience, Chapter 2 introduces basic concepts very briefly and formally, sometimes in non-standard ways without examples (e.g., transactions as tuples), formally incorrect (e.g., containment of an itemset in a transaction), or not at all (e.g., the notion of \(k\)-itemset). Moreover, some descriptions of algorithms are incomplete (e.g., computation of new frontier set in the Apriori-algorithm is missing) while others are too short to be of much value (e.g., FP-growth, OPUS). While later chapters do not suffer from such shortcomings, they have a strong bias towards the authors' own, previously unpublished work, but do not present an overview of the field, e.g., neither mining of multi-level or multi-dimensional rules nor the impact of constraints are addressed. To sum up, I cannot recommend this book as an introductory text and I advise anyone more involved to check the Contents whether the book addresses their interests.