Browse Books

Reviewer: Immo O. Kerner

The book is divided into two parts, each serving a different but complementary purpose. The first part is an introduction to machine learning, with special reference to the development of expert systems. It surveys the major systems (Perceptron, Winston's Program, Interactive Dichotomizer ID3, AQ11, Meta-Dendral, Bacon 4, and BEAGLE) describes how they work and how they are put to practical work, and considers future prospects. Within five chapters, the first author provides an introduction to machine learning, describing black box methods, learning structural descriptions, and evolutionary strategies, and he presents an outlook towards the learning machine. The objective of the second part is to demonstrate machine learning in action within an important contemporary field of information retrieval. After a short introduction, the second author gives an overview of existing work: language of information retrieval, the role of learning, and knowledge-sparse and knowledge-rich learning. This is followed by a discussion on applications in the field of medicine, with a detailed case study description of the Medlars System that fills about 80 pages. The book is illustrated by many diagrams and contains a glossary of over 100 technical terms. Each chapter in each part is follwed by a list of references. Also, some exercises are included in Part 2. This book is recommended for students, teachers, researchers, and workers in AI, cognitive sciences, information retrieval, machine learning, and psychology. The book is indeed much more than an advertisement for the system developed by the authors.