Abstract

A new algorithm for Support Vector regression is proposed. For a priori chosen , it automatically adjusts a flexible tube of minimal radius to the data such that at most a fraction of the data points lie outside. The algorithm is analysed theoretically and experimentally. 1 Introduction Support Vector (SV) machines comprise a new class of learning algorithms, motivated by results of statistical learning theory [4]. Originally developed for pattern recognition, they represent the decision boundary in terms of a typically small subset [2] of all training examples, called the Support Vectors. In order for this property to carry over to the case of SV Regression, Vapnik devised the so-called "-insensitive loss function [4] jy \Gamma f(x)j " = maxf0; jy \Gamma f(x)j \Gamma "g, which does not penalize errors below some " ? 0, chosen a priori. His algorithm, which we will henceforth call "-SVR, seeks to estimate functions f(x) = (w \Delta x) + b; w; x 2 R N ; b 2 R; (1) based on data (x...

Citations