一种基于圆形分布的支撑向量机核选择方法

A Kernel Selection Approach of Support Vector Machine Based on Characteristics of Circle Distribution

针对目前支撑向量机核函数的选择没有统一规则的现状,提出了一种结合数据分布特征进行支撑向量机核选择的方法。首先,采用多维尺度分析方法对高维数据集合理降维,提出判断数据集是否呈圆形分布的算法,在得到数据集分布特征的基础上进行核选择,达到结合数据分布特征合理选择支撑向量机核函数的目的。实验结果表明:呈圆形分布的数据集采用极坐标核进行分类,识别率达到100%,训练时间短,优于采用神经网络、决策树、高斯核及多项式核的分类效果。该方法提高了支撑向量机的泛化能力。

The kernel selection has no unified rules for support vector machine( SVM). Based on the characteristics of dataset distribution,a new way to select the kernel function was presented. First dimension reduction of the high dimensional dataset was processed with MDS method. Then an algorithm was put forward to judge whether dataset is circle distribution or not. On the basis of determining circle distribution,how to select the kernel function was discussed to achieve the purpose of selecting SVM kernel function with data distribution characteristics. The experimental results illustrate that the classification recognition rate of circle datasets reaches 100% with polar kernel and the training time is the shortest. The classification effect is better than that of using neural network,decision tree,gaussian kernel SVM and polynomial kernel SVM. The method can improve the generalization ability of SVM.