A new feature selection method based on a validity index of feature subset

• Published in: Pattern recognition letters Vol. 92; pp. 1 - 8
• Main Authors: Liu, Chuan, Wang, Wenyong, Zhao, Qiang, Shen, Xiaoming, Konan, Martin
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.06.2017; Elsevier Science Ltd
• Subjects: Approximation; Classification; Datasets; Feature selection; Filter methods; Filtering systems; Pattern recognition; Search algorithms; Selection; Support vector machines; Wrapper methods; Wrapper methods; Filter methods; Feature selection
• ISSN: 0167-8655; 1872-7344
• DOI: 10.1016/j.patrec.2017.03.018

•A new statistical LW-index for labeled feature set is proposed.•An new filter algorithm, i.e. SFS-LW, is presented.•It can obtain similar classification accuracy as the wrapper methods.•It is nearly ten times faster than the wrapper methods. The wrapper feature selection method can achieve high classification accuracy. However, the cross-validation scheme of the wrapper method in evaluation phase is very expensive regarding computing resource consumption. In this paper, we propose a new statistical measure named as LW-index which could replace the expensive cross-validation scheme to evaluate the feature subset. Then, a new feature selection method, which is the combination of the proposed LW-index with Sequence Forward Search algorithm (SFS-LW), is presented in this paper. Further, we show through plenty of experiments conducted on nine UCI datasets that the proposed method can obtain similar classification accuracy as the wrapper method with centroid-based classifier or support vector machine, and its computation cost is approximate to the compared filter methods.