摘要
对高斯核函数支撑向量机受参数σ影响较大的缺陷,提出了一种基于规范群算法的混合分类算法———基于规范群的混合支撑向量机,通过仿真结果表明了该算法的正确性和有效性.
The selection of parameter has great effect on the gauss kernel-based support vector. In order to overcome the defect, a new hybrid classification algorithm combining gauge group with gauss kernel algorithm is presented. The simulation result indicates that the hybrid algorithm is of accuracy and validity.
出处
《沈阳理工大学学报》
CAS
2005年第1期6-10,共5页
Journal of Shenyang Ligong University
关键词
支撑向量机
核函数
规范群
support vector machine
kernel function
gauge group
