Feature Selection Using Cubic Smoothing Spline and Robust Regression

Abstract

An efficient feature selection approach based on the combination of cubic smoothing spline and robust regression is presented for classification applications in this study. Six different data sets are used to test the proposed feature selection algorithm. The success of proposed algorithm is evaluated by using K-Nearest Neighbor (KNN) algorithm and Discriminant analysis. Obtained simulation results show that proposed feature selection approach has high classification accuracy rate with fewer number of features.