Introduction to Machine Learning (2ed), by Ethem Alpaydin, MIT Press, 2010. ISBN 0-262-01243-X.
This book provides students, researchers, and developers a comprehensive introduction to the machine learning techniques. It is structured primarily as coursebook, which is a valuable teaching textbook for graduates or undergraduates. This book is also a good resources for self-study by researches and developers, but they have to be familiar with AI and advanced mathematics.
This book begins with an introduction chapter, followed by 18 chapters plus an appendix. Each chapter presents a stand-alone topic, beginning with a brief introduction and ending with notes. Therefore, the readers can quickly obtain an overview for the topic and catch the possible direction to further development in this subject area. The book covers a variety of machine learning techniques: supervised and unsupervised learning, parametric and nonparametric methods. All of these are followed by methods of how to assess and compare classification algorithms, combine multiple learners, and reinforce learning procedure.
As a book dealing with machine learning, it presents a varied collection of the different programming techniques; however some topics are not well organized. In Bayesian decision theory, the multiple inputs and outputs case is a complex area, but this chapter only briefly mentions the preliminary knowledge. In chapter 3, Naive Bayes classifier is also a really important classifier, but it is not given the amount of space that it deserves. Chapter 6 has a section on linear discriminate analysis (6.6), which is better to located in Chapter 10.
Generally, examples present at the end of each chapter are often too simple. They lack some mathematical problems which can only appear when the size of examples is relative large. Furthermore, for the list of reference at the end of each chapter, it is disappointing that almost all references are dated from 2000 or earlier, but the publication of this book is 2010.
(Mitchell, 1997) is the closest substitute for this book, and worth keeping in your library. (Russell and Norvig, 2003) covers important aspects of machine learning as well as many related concepts such as knowledge representation, and different search heuristics. Although much of this content has earlier been covered by (Mitchell, 1997) and less so by (Russell and Norvig, 2003), this book still stands out.
- Mitchell, T. M. (1997). Machine Learning. McGraw-Hill, New York.
- Russell, S. J. and Norvig, P. (2003). Artificial Intelligence: A Modern Approach (2ed). Prentice Hall.