CAJ | 학술논문

本文提出了基于支持向量回归机(SVR)的一种新分类算法.它和标准的支持向量机(SVM)不同:标准的支持向量机(SVM)采用固定的模度量间隔且最优化问题与参数有关.本文中我们可以用任意模度量间隔,得到的最优化问题是无参数的线性规划问题,避免了参数选择.数值试验表明了该算法的有效性.
본문제출료기우지지향량회귀궤(SVR)적일충신분류산법.타화표준적지지향량궤(SVM)불동:표준적지지향량궤(SVM)채용고정적모도량간격차최우화문제여삼수유관.본문중아문가이용임의모도량간격,득도적최우화문제시무삼수적선성규화문제,피면료삼수선택.수치시험표명료해산법적유효성.
A new algorithm to solve classification problems is obtained based on the support vector regression (SVR). It is different form the standard Support vector machine (SVM): The margin is measured by fixed norm and the optimization problems depend on the parameters in the standard SVM. In this paper, we can measure the margin by arbitrary norm, and the deduced optimization problem is a linear programming without parameters. Preliminary experiments also show the validity of our new algorithm.