A new feature selection method based on relief and SVM-RFE

In original data, there may exist redundant features, irrelevant features, noisy features besides informative features. Extracting informative features while eliminating the others is the goal of feature selection. This paper proposed a new feature selection algorithm based on Relief algorithm and SVM-RFE algorithm, and it is strongly targeted to eliminate the unnecessary features. Finally, We test the proposed method on three data sets from UCI, and treat accuracy, size of optimal subset, time-cost as evaluations, the experimental results show that the proposed algorithm has a better performance than Relief algorithm and SVM-RFE algorithm except time-cost.