用于回归的临近支持向量机  

Proximal Support Vector Machine for Regression

作　　者：杜喆[1] 胡廷锋[2] 刘三阳[1] 

机构地区：[1]西安电子科技大学应用数学系,西安710071 [2]洛阳师范学院数学科学学院,洛阳471022

出　　处：《计算机科学》2009年第1期126-127,157,共3页Computer Science

基　　金：国家自然科学基金项目(60574075,60705004)资助

摘　　要：将临近支持向量分类机应用在回归问题上,提出临近支持向量回归机,给出线性与非线性情况下的回归函数,该方法比支持向量回归机(SVR)问题减少了参数和一半变量,比最小二乘支持向量回归机(LSSVMR)求解公式更加简单,且核函数不需要满足Mercer条件。数值实验结果表明,与SVR和LSSVMR相比,该方法的学习速度更快,且泛化能力较之不相上下。A new technique-proximal support vector regression machine （PSVRM） was obtained when extended proximal support vector machine classifier towards regression problem, the linear and nonlinear regression function of PSVRM was proposed. PSVRM has less arguments and half variables than support vector regression（SVR）, the solving formula is much simpler than that of least square support vector machine for regression（LSSVMR） ,and the kernel is no need to satisfy the Mercer＇s condition. Numerical simulations show that, comparing with SVR and LSSVMR, PSVRM is faster and the generalization ability is sometimes faster than the former two.

关 键 词：临近支持向量机 回归 支持向量机 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]