Machine Learning and Statistics

the long version

Machine Learning and Statistics is a result of the authors' participation in the 1994 European Conference in Machine Learning. This important collection of contributions was adapted from conference workshop material and reworked to address readers of diverse backgrounds and skills. For newcomers to the field, a thorough introduction surveys the various topics and supplies numerous references for further reading. The book's main focus is on classification, the most common area of intersection. The classification process uses information about a new example to assign the example to one of a known number of classes. Such methods typically involve a rule learned from an initial set of data, which is where ML comes into play. Other topics covered include prediction, control, and an introduction to methods' of knowledge discovery in databases - a skill that has become especially relevant with the explosion in large-scale databases. Timely, practical, and innovative, this book offers a number of new algorithms and draws on real-world examples including financial and medical applications. It also includes two chapters on loans/credit applications that help identify bad risks and good customers - useful for those working with credit scoring and bad debt analysis.