Feature selection for Support Vector Machines via Mixed Integer Linear Programming

作     者：Maldonado, Sebastian Perez, Juan Weber, Richard Labbe, Martine 

作者机构：Univ Los Andes Santiago Chile Univ Chile Dept Ind Engn Santiago Chile Univ Libre Bruxelles Dept Comp Sci B-1050 Brussels Belgium 

出 版 物：《INFORMATION SCIENCES》 (信息科学)

年 卷 期：2014年第279卷

页      面：163-175页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Institute of Complex Engineering Systems [ICM: P-05-004-F, CONICYT: FBO16] FONDECYT project Interuniversity Attraction Poles Programme 

主　　题：Feature selection Support Vector Machine Mixed Integer Linear Programming 

摘      要：The performance of classification methods, such as Support Vector Machines, depends heavily on the proper choice of the feature set used to construct the classifier. Feature selection is an NP-hard problem that has been studied extensively in the literature. Most strategies propose the elimination of features independently of classifier construction by exploiting statistical properties of each of the variables, or via greedy search. All such strategies are heuristic by nature. In this work we propose two different Mixed Integer Linear Programming formulations based on extensions of Support Vector Machines to overcome these shortcomings. The proposed approaches perform variable selection simultaneously with classifier construction using optimization models. We ran experiments on real-world benchmark datasets, comparing our approaches with well-known feature selection techniques and obtained better predictions with consistently fewer relevant features. (C) 2014 Elsevier Inc. All rights reserved.