支持向量机核函数选择研究与仿真

J4 ›› 2015, Vol. 37 ›› Issue (06): 1135-1141.

支持向量机核函数选择研究与仿真

梁礼明，钟震，陈召阳

(江西理工大学电气工程与自动化学院，江西赣州 341000)

收稿日期:2014-05-18 修回日期:2014-09-18 出版日期:2015-06-25 发布日期:2015-06-25
基金资助:
国家自然科学基金资助项目(5136501)；江西省自然科学基金资助项目(20132BAB203020)；江西省教育厅科学技术研究项目(GJJ13430)

Research and simulation of kernel function
selection for support vector machine

LIANG Liming，ZHONG Zhen，CHEN Zhaoyang

(School of Electrical Engineering and Automation,Jiangxi University of Science and Technology,Ganzhou 341000,China)

Received:2014-05-18 Revised:2014-09-18 Online:2015-06-25 Published:2015-06-25

摘要/Abstract

摘要：

支持向量机是一种基于核的学习方法，核函数选取对支持向量机性能有着重要的影响，如何有效地进行核函数选择是支持向量机研究领域的一个重要问题。目前大多数核选择方法不考虑数据的分布特征，没有充分利用隐含在数据中的先验信息。为此，引入能量熵概念，借助超球体描述和核函数蕴藏的度量特征，提出一种基于样本分布能量熵的支持向量机核函数选择方法，以提高SVM学习能力和泛化能力。数值实例仿真验证表明了该方法的可行性和有效性。

关键词: 支持向量机；核函数；样本分布；先验信息；能量熵

Abstract:

Support vector machine (SVM)is a kind of learning method based on kernel. Kernel function selection has significant influence on the performance of SVM, so how to effectively select the kernel function is an important problem in the field of SVM research. At present most of the kernel function selection methods neither consider the characteristics of data distribution, nor make full use of the implicit prior information in the data. We introduce a concept of energy entropy; along with the super sphere description and measurement features of kernel function, we put forth a method of kernel function selection based on the energy entropy of sample distribution so as to improve the learning ability and generalization ability of SVM. Simulations on numerical examples show that the method is feasible and effective.

Key words: SVM;kernel function；sample distribution；prior information；energy entropy

梁礼明，钟震，陈召阳. 支持向量机核函数选择研究与仿真[J]. J4, 2015, 37(06): 1135-1141.

LIANG Liming，ZHONG Zhen，CHEN Zhaoyang. Research and simulation of kernel function
selection for support vector machine [J]. J4, 2015, 37(06): 1135-1141.