Wang L. (Ed.) Support Vector Machines: Theory and Applications

Berlin: Springer, 2005. - 434 p.
This volume is composed of 20 chapters selected from the recent myriad of novel SVM applications, powerful SVM algorithms, as well as enlightening theoretical analysis. Written by experts in their respective fields, the first 12 chapters concentrate on SVM theory, whereas the subsequent 8 chapters emphasize practical applications, although the decision boundary separating these two categories is rather fuzzy.
Kecman first presents an introduction on the SVM, explaining the basic theory and implementation aspects. In the chapter contributed by Ma and Cherkassky, a novel approach to nonlinear classification using a collection of several simple (linear) classifiers is proposed based on a new formulation of the learning problem called multiple model estimation. Pelckmans, Goethals, De Brabanter, Suykens, and De Moor describe componentwise Least Squares Support Vector Machines (LS-SVMs) for the estimation of additive models consisting of a sum of nonlinear components.