Abstract:
This paper researches the selection problem of kernel function for Relevance Vector Machine(RVM). Improved Gauss kernel function is proposed. The characteristic of improved Gauss kernel function and normal Gauss kernel function are compared. The improving performance of proposed kernel function is validated. Besides the improving of single kernel function, multi-kernel RVM is researched, by combining local Gaussian kernel and global polynomial kernel, form multi-kernel function, and use it in RVM. Comparison experiments of kinds of kernel functions run on different datasets, and the performance of improved Gauss kernel function and mixture kernel function are validated.
Key words:
Relevance Vector Machine(RVM),
improved Gauss kernel function,
multi-kernel
摘要: 针对相关向量机中的核函数选择问题进行研究,对高斯核函数进行改进,提出修正的高斯核函数方法,并比较改进的高斯核函数与普通高斯核函数的特性,证明提出的核函数的优良特性。在对单一核函数改进的基础上,进行多核相关向量机核函数的研究,结合局部性高斯核函数和全局性多项式核函数形成混合核函数,并运用于相关向量机。在不同大小的数据集上对几种核函数进行对比实验,验证修正的高斯核函数及混合核函数的性能。
关键词:
相关向量机,
修正的高斯核函数,
多核
CLC Number:
YANG Liu, ZHANG Lei, ZHANG Shao-Xun, LIU Jian-Wei. Comparison Research of Single Kernel and Multi-kernel Relevance Vector Machine[J]. Computer Engineering, 2010, 36(12): 195-197.
杨柳, 张磊, 张少勋, 刘建伟. 单核和多核相关向量机的比较研究[J]. 计算机工程, 2010, 36(12): 195-197.