This book is directed to machine learning (ML) students, researchers, and practitioners searching for a better understanding of different concepts and techniques adopted in data and knowledge manipulation and classification. It is especially useful for those looking for better methods, better solutions, and new ways of applying combined classifiers, multiple classifiers, ensemble classifiers, classifier fusion, data stream classifiers, and hybrid machine learning methods. All of these methods are used for data and knowledge integration, transformation, classification, evaluation, and use. The author presents an up-to-date review of recent advances in this area.

Absolute beginners are not the intended audience of the book. In fact, it may be quite difficult to use it as an introductory text on machine learning concepts for those unfamiliar with the topic. On the other hand, the author presents a quite complete and interesting review and overview of important topics related to machine learning and classifiers in chapter 1.

Chapter 1 covers such concepts as: the curse of dimensionality, feature selection, and feature reduction; types of features (discrete, discretized, continuous values); costs of classification, training, testing, and misclassification; bias variance, overfitting, and the model selection problem; learning evaluation, generalization, and cross-validation; the no free lunch theorem; Occam’s razor; post-pruning; minimum description length (MDL); VC dimension; and batch, incremental, online, and interactive learning. This chapter does not focus on a single ML approach, but presents different methods and knowledge acquisition/representation techniques, such as Bayesian classifiers, k-nearest neighbors, rule-based classifiers, artificial neural network classifiers, support vector machines (SVMs), decision trees, and hybrid classifiers. A special focus is always put on the evaluation methods since these are very important when building hybrid systems.

Chapter 2 is focused on data and knowledge hybridization. The author discusses data and knowledge quality, merge/unification, transformation, and consistency. This chapter considers different knowledge and data representation methods, as well as partitioning and data privacy aspects.

In chapter 3, the author explores classifiers and how to combine them, covering topology, ensembles, diversification/complementarity, and combination rules.

Finally, chapter 4 presents hybrid classifiers from the application point of view, and covers one-class classification, the absence of counterexamples, imbalanced data classification, data stream classification, and concept drift and related issues.

In conclusion, this is a very interesting, complete, and up-to-date book about various aspects of machine learning and decision making using hybrid classifiers. Although the author makes this book accessible to students and practitioners, it is probably more oriented to advanced undergraduate or graduate courses focused on improving machine learning methods and applications.