A SVM Kernel Parameter Selection Approach Based on the Gauss Distribution

孔晓斌; 曹棣; 张素兰

您当前的位置：

首页 >

文章列表页 >

A SVM Kernel Parameter Selection Approach Based on the Gauss Distribution

更新时间：2026-09-14

- A SVM Kernel Parameter Selection Approach Based on the Gauss Distribution
- 一种基于高斯分布的SVM核参数选择方法
- Issue 4, Pages: 343-346(2008)
- 作者机构：
  
  1. 太原科技大学计算机科学与技术学院
  2. 太原理工大学计算机与软件学院
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2008
- 稿件说明：
移动端阅览
孔晓斌, 曹棣, 张素兰. A SVM Kernel Parameter Selection Approach Based on the Gauss Distribution[J]. 2008, (4): 343-346.
DOI：

孔晓斌, 曹棣, 张素兰. A SVM Kernel Parameter Selection Approach Based on the Gauss Distribution[J]. 2008, (4): 343-346. DOI：

摘要

针对支持向量机(Support Vector Machine

SVM)研究中的核选择问题

在Amari和Wu提出的保形变换方法的基础上

提出了一种基于高斯分布的SVM核参数选择方法.分析了确定数据分布特征的重要性

给出了判断数据呈高斯分布的方法

探讨了SVM核函数及其参数选择与数据分布的相关性.在Matlab实验环境下

采用两组数据集进行了数值仿真

仿真结果说明了本文所提方法的可行性与有效性.

Abstract

Aiming at the problem of kernel selection for support vector machine(SVM)

a new method for selecting the kernel parameter was presented.It was based on the Gauss distribution and conformal transformation method presented by Amari and Wu.The importance of determining the characteristics of data distribution was analyzed

and an approach to determine the Gauss distribution was presented.The correlation of data distribution and how to select the kernel parameter was discussed.Two groups of data sets were used to carry on the numerical simulation by Matlab.The simulating results demonstrate the feasibility and the effectiveness of the presented approach.

关键词

Keywords

references

Vapnik V.The Nature of Statistical Learning Theory[M].Springer2Verlag,New York,Inc,2000.

Burges J C.A Tutorial on Support Vector Machines for Pattern Recognition[M].Kluwer Academic Publishers,Boston,1999.

Wang W J,Xu Z B,Lu W Z,et al.Determination of the spread parameter in the Gaussian kernel for classificationand regression[J].Neurocomputing,2003,55(3-4):643-663.

Duan K B,Keethi S,Poo A N.Evaluation of simple performance measure for tuning SVM hyperparameters[J].Neurocomputing,2003,51:41-59.

Cherkassky V,Ma Y Q.Practical selection of SVM parameters and noise estimation for SVM regression[J].NeuralNetworks,2004,17:113-126.

Golda C,Sollichb P.Modle selection for support vector machine classification[J].Neurocomputing,2003(55):221-249.

Gavin C,Nicola T.Efficient leave-one-out cross-validtion of kernel fisher discriminant classifiers[J].PatternRecognition,2003,11:2585-2592.

Amaris S,Wu S.Improving support vector machine classifiers by modifying kernel functions[J].Neural Networks,1999(12):783-789.

赵倩,李宏伟.一种产生广义高斯分布随机数的算法[J].应用数学,2004,17(增1):64-69.Zhao Qian,Li Hongwei.A new algorithm for generating generalized gaussian random number[J].MathematicaApplicata,2004,17(s1):64-69.(in Chinese)

Walden A T.NonGaussian,reflectivity,entropy and reconvolution[J].Geophysics,1985,50(12):2862-2888.

Stacy E W.A generalization of gamma distribution[J].Ann Math Stat,1962,28:1187-1192.

Mahesh K V,Behnaam A.Parametric generalized gaussi-an density estimation[J].Journal Acoustical Society ofAmerican,1989,86(4):1404-1415.

Morris H D.Probability and Statistics[M].Addisonwelsy Publishing Company,1975.

Views

328

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

基于特征融合与SVM的内镜图像分类算法研究

Research on Continuous Identity Authentication Based on Mouse and Window Behavior

A Support Vector Machine Integration Method Based on Nearest Neighbor Classification Accuracy

改进混合特征模型聚类的文本情感分类算法研究

模拟电路故障诊断的神经网络模型研究

Related Author

曹聪

侯木舟

吕志贤

金海彬

宋礼鹏

田杰

吕晓燕

陈立潮

Related Institution

中南大学数学与统计学院

中北大学大数据学院

山西医科大学计算机教学部

甘肃联合大学电子信息工程学院

中北大学信息与通信工程学院

⁰