求解非半正定核Huber–支持向量回归机问题的序列-控制理论与应用.PDFVIP

27 9 Vol. 27 No. 9 2010 9 Control Theory Applications Sep. 2010 :2010 Huber– , , , , (1. , 210094; 2. , 214122) : (SMO)(SVM) . , . , SMO, Huber-SVR . Huber-SVR, . : ; ; ; Huber– : TP18 : A Sequential-minimal-optimization algorithm for solving Huber-suppor-vector-regression with non-positive semi-definite kernels ZHOU Xiao-jian , MA Yi-zhong , ZHU Jia-gang , LIU Li-ping , WANG Jian-jun (1. Department of Mangement Science and Technology, Nanjing University of Science and Technology, Nanjing Jiangsu 210094, China; 2. School of Information Engineering, Jiangnan University, Wuxi Jiangsu 214122, China) Abstract: Sequential-minimal-optimization(SMO) algorithm is effective in solving large-scale support-vector- machine(SVM) problems. However, the existing algorithms require the kernel functions to be positive definite(PD) or positive semi-definite(PSD), thus limiting their applications. Having considered their deficiencies, we propose a new al- gorithm for solving Huber-SVR problems with non-positive semi-definite(non-PSD) kernels. This algorithm provides desirable regression-accuracies while ensuring the convergence. Thus, it is of theoretical and practical significance. Key words: support-vector-machine; non-positive semi-definite kernel; sequential-minimal-optimization algorithm; Huber-support vector regression 1 (Introduction) , [2] (SVC) , SVM , [3] SVR. , (MINOS,