基于PSO优化的SVM预测应用研究.pdfVIP

第 26卷第 3期 计 算 机 应 用 研 究 V o.l 26 N o. 3 2009年 3月 ApplicationR esearch of Computers Mar. 2009 PSO SVM * , (东北林业大学信息与计算机工程学院, 哈尔滨 150040) : 支持向量机参数对支持向量机的性能有着重 影响,参数选择问题是支持向量机的重 研究内容针 对此问题,提出一种基于粒子群优化算法的支持向量机参数选择方法实验结果表明, 经粒子群优化算法优化 的支持向量机回归模型具有较高的预测精度, 粒子群优化算法是选取支持向量机参数的有效方法 : 支持向量机; 粒子群优化算法; 握钉力; 预测 : TP181 : A : 2009) upport vector machine optmi ized by particle swarm optmi ization algorithm for holding nail force forecasting REN H onge, HUO M andong (College of Inf ormation ComputerEng ineering, N ortheast Forestry University, H arbin 150040, China) Ab stract: In the param eters of support vectormachine( VM ) have mi portant effect to VM perform ance. The param eters se lection is the mi portant research content of the VM. To th is problem, this paper proposed one kind of m ethod to choose the param eters of the VM by particle sw arm optmi ization algorithm ( P O ). The expermi ent result ind icates the VM regression m odel optmi ized by P O have h igh forecast accuracy, and P O is one kind of effectivem ethod for VM param eters choosing. K ey words: VM ( support vectorm ach ine); P O( particle sw arm optmi ization algorithm); holding nail force; forecast , P O VM , 0 ( least square support vec torm achine, L VM ) , , , , 1 L VM P O [ 1] , 11 L VM , ,