This book presents a comprehensive introduction to Support Vector Machine which is one of the most popular and effective classification method used in learning theory.
The method is based on the possibility of extending the classical linear classifier methods, by being linear in representation spaces of much larger dimensionality. This can be achieved without really computing in these larger spaces by using Mercer kernels.
After an introduction concerning learning methodology, a chapter on linear learning introduce the basis of the algorithmic apparatus in simple conditions. Chapter 3 shows how the method can be extended using kernels in “feature spaces”. Chapter 4 presents the main results of Vapnik Chervonenkis theory that make a basis to generalization theory. Chapter 5 investigate optimization theory in that it provides efficient methods to compute the minimizations involved. Chapter 6 is devoted to support vector classification in its various modalities. Chapter 7 treats of the implementation of the method. Finally, Chapter 8 provides applications examples.
This is an excellent book that presents a complete and readable without big requirements in mathematical functional analysis introduction to a very important classification scheme.