基于粗糙集粒子群支持向量机的特征选择方法

Feature Selection Based on RS-PSO-SVM

摘要: 将Filter型粗糙集属性约简方法与PSO-SVM方法相结合,提出一种新的粗糙集粒子群支持向量机(RS-PSO-SVM)特征选择方法.给出了该方法的特征选择具体步骤,并对比分析了所提方法的性能.仿真实验表明:提出的RS-PSO-SVM特征选择方法是有效的,在保证所选特征集为最优情况下,极大地缩短所用时间,可以将其应用在多维数据的特征选择中.

Abstract: A new RS-PSO-SVM feature selection combined Filter Rough Set attribute reduction with PSO-SVM was proposed. The specific steps of the new method were given, and the performance was analyzed contrasted with other two methods. Simulation experiments show that, the proposed feature selection method is effective which can reduce time consumed largely and ensure the selected feature set is the best. The proposed method can be used in the feature selection of multidimensional data.