Randomized Kernel Methods for Least-Squares Support Vector Machines

摘要

The least-squares support vector machine is a frequently used kernel methodfor non-linear regression and classification tasks. Here we discuss severalapproximation algorithms for the least-squares support vector machineclassifier. The proposed methods are based on randomized block kernel matrices,and we show that they provide good accuracy and reliable scaling formulti-class classification problems with relatively large data sets. Also, wepresent several numerical experiments that illustrate the practicalapplicability of the proposed methods.