The research of SMO algorithm self-adaption improvement on SVM

Abstract

The SVM (Support Vector Machine) is a kind of important statistical machine learning algorithm. The SMO (Sequential Minimal Optimization) is one of the algorithms on SVM. It is more effective method in practical application. And the SMO algorithm is the solution of support vector machine quadratic programming problem for a series of smaller problems decomposition, thus it realizes serial minimum optimization. The method is used in SMO algorithm of adaptive learning thoughts, and is solving convex quadratic programming optimization problems on the basis on improvement, which has been proposed in this paper. Therefore, based on the idea of adaptive learning algorithm is improved the SMO. The SVM algorithm can adapt to the practical application of more rapid and efficient needs.