基于改进SVM-RFE的特征选择方法研究

SVM-RFE特征选择算法是一种有效的特征选择方法,具有较高的应用价值。针对传统SVM-RFE特征选择算法中SVM参数(γ和C)难以确定的问题,本文采用粒子群算法搜索SVM的参数。然后将特征向量映射到SVM参数γ确定的核空间中并进行特征选择,有效地将特征选择与SVM分类器设计关联起来。仿真结果表明,特征选择后的数据集仍能保证SVM分类器具有较高的分类正确率。

Research on Feature Selection Method Based on Improved SVM- RFE

SVM-RFE feature selection arithmetic,which is an effective method for feature selection,has a rather high value for application.Parameters of SVM in the feature selection arithmetic based on SVM-RFE are difficult to be determined.It is proposed to use the PSO arithmetic to search reasonable SVM parameters.Then feature vector was mapped to the kernel space determined by the SVM parameter and feature selection was operated.This strategy associated design of SVM classifier with feature selection efficiently.Si...