Machine Learning

Citation

(1998), "Machine Learning", Industrial Robot, Vol. 25 No. 6. https://doi.org/10.1108/ir.1998.04925fae.001

Machine Learning

Machine Learning

T.M. MitchellMcGraw-Hill1997414 pp.ISBN 0-07-115467-1£21.99 (paperback)

Based on recent successes in the field of machine learning, the goal of the book is to present the key algorithms and theory that form the core of the subject. The author draws on concepts and results from the many fields contributing to the field from mathematics, knowledge engineering and cognitive science, viewing the subject from a number of diverse perspectives in an attempt to understand the problems, algorithms and assumptions that underlie each approach. The text is a broad-based introduction to the topic of machine learning, drawing on the major fields that contribute to the subject.

The book is logically split along chapter lines, each offering an explanation of a different technique in learning technology, after two chapters of a general introduction to the topic. The remaining 11 specialised chapters are capable of being read out of sequence if desired, each offering a self-contained introduction and explanation of the main points in the particular subject area. The subject areas covered include: tree learning, artificial neural networks, statistics and information theory, Bayesian analysis and computational and instance-based learning methods. Further chapters cover genetic algorithms, rules sets, explanation-based learning, the combination of prior knowledge with training data and reinforcement learning. The end of each chapter contains a summary of the main concepts covered, suggestions for further reading, exercises and references. Additional updates to chapters as well as data sets and implementations of algorithms are available from the internet at http://www.cs.cmu.edu/~tom/mlbook.html; specifically the examples of neural network face recognition, Bayesian learning for classifying news text articles and code for a decision tree implementation. In addition the author has made available teaching materials including slides and lecture notes for a course based on the book, all resources being free to download from the above page.

The book itself is a welcome addition to the field, offering a very easily read introduction to a fast moving and broad field of research. The author has made the material approachable for readers from most backgrounds, while still maintaining a degree of rigour in definitions. As with all texts of this nature, it is not the last word in reference material, but makes an excellent and interesting introduction to the diversity of subjects that make up machine learning.