Presents a comprehensive and self-contained introduction to Gaussian processes, which provide a principled, practical, probabilistic approach to learning in kernel machines. This book deals with the supervised-learning problem for both regression and classification, and includes algorithms. It also contains illustrative examples and exercises.