一种新颖的基于混合不确定性的特征选择方法

A novel feature selection method based on mixed uncertainty

基于混合不确定性的特征选择方法(简称SU-P方法),利用对称不确定性找出相关特征,并利用偏相关分析去除数据集中的冗余特征。将SU-P方法与其它传统算法进行了仿真对比。

The mixed-uncertainty-based feature selection method(SU-P method in short),is to use symmetric uncertainty to find the relevant features and partial correlation analysis to remove redundant features in the data set.It is compared with other traditional algorithms in simulation.