Un algoritmo para el entrenamiento de máquinas de vector soporte para regresión

Abstract

The aim of the present paper is twofold. Firstly an introduction to the ideas of Support Vector regression is given. then a new and simple algorithm, suggested by the work of Campbell y Cristianini in [16], is proposed which solves the corresponding quadratic programming problem in an easy fashion. The algorithm is illustrated by example and compared with classical regression. Keywords: Support vector Machines, E-Support Vector regression.