Working Set Selection Using Second Order Information for Training Support Vector Machines

TLDR

Working set selection is an important step in decomposition methods for training support vector machines (SVMs). This paper develops a new technique for working set selection in SMO‑type decomposition methods. The method uses second‑order information to achieve fast convergence. The method achieves linear convergence theoretically and is experimentally faster than first‑order selection methods.

Abstract

Working set selection is an important step in decomposition methods for training support vector machines (SVMs). This paper develops a new technique for working set selection in SMO-type decomposition methods. It uses second order information to achieve fast convergence. Theoretical properties such as linear convergence are established. Experiments demonstrate that the proposed method is faster than existing selection methods using first order information.