This book is devoted to pattern recognition and machine learning — advanced areas of software science. It is indeed the first text to provide a unified, self-contained introduction to visual pattern recognition and machine learning as it is based on the idea that “the generation and transformation of information representations can be used to explain both problems”.
The book explains basic concepts and algorithms for representative methods of both pattern recognition and learning by computer. It is also useful as a general introduction to artificial intelligence and knowledge engineering. It can be a very useful textbook for students who are going to study software science. The book covers important research topics and gives many examples. The text is exceptionally well-structured and provides a wealth of information regarding both the theory and the practice of recognition and learning by computer. The book is divided into 10 chapters: (1) Recognition and learning by a computer, (2) Representing information, (3) Generation and transformation of representations, (4) Pattern feature extraction, (5) Pattern understanding methods, (6) Learning concepts, (7) Learning procedures, (8) Learning based on logic, (9) Learning by classification and discovery, (10) Learning by neural networks. Each chapter includes a summary, keywords and exercises. The text is complemented by a preface and a study guide which help the readers understand the philosophy and the structure of the book. An appendix including sample programs for popular learning algorithms using neural networks, answers to exercises which include additional representation methods and algorithms, a bibliography and a very good subject index. The bibliography includes referenced books, titles for additional reading, academic journals, research papers and collections of essays.
The book describes 43 algorithms with a particularly effective use of diagrams. In addition, the text is accompanied by sample programs in \(C\), Lisp, Pascal and Prolog. The last chapter and the appendix address the very recent neural network technique which is beginning to show a great potential for high performance in pattern recognition and learning by computer.
The exercises at the end of each chapter are particularly useful since they encourage the development of software and provide a deeper understanding of the text.
Overall, the author has the merit of having written a very interesting, up-to-date and practical book.