Introduction

The term machine learning covers a broad range of computer programs. In general, any computer program that can improve its performance at some task through experience (or training) can be called a learning program [1]. There are two general types of learning: inductive, and deductive. Inductive learning aims to obtain or discover general rules/facts from particular training examples, while deductive learning attempts to use a set of known rules/facts to derive hypotheses that fit the observed training data. Because of its commercial values and variety of applications, inductive machine learning has been the focus of considerable researches for decades, and most machine learning techniques in the literature fall into the inductive learning category. In this book, unless otherwise notified, the term machine learning will be used to denote inductive learning.