Canonical duality solution to support vector machine

Yubo Yuan, Feilong Cao

Abstract: Support vector machine (SVM) is one of the most popular machine learning method and educed from a binary data classification problem. In this paper, a new duality theory named canonical duality theory is presented to solve the normal model of SVM. Several examples are illustrated to show that the exact solution can be obtained after the canonical duality problem being solved. Moreover, the support vectors can be located by non-zero elements of the canonical dual solution.