Fast SVM classifier for large-scale classification problems

Wang，Huajun1; Li，Genghui2; Wang，Zhenkun3

2023-09-01

10.1016/j.ins.2023.119136

Information Sciences   影响因子和分区

0020-0255

1872-6291

Support vector machines (SVM), as one of effective and popular classification tools, have been widely applied in various fields. However, they may incur prohibitive computational costs when solving large-scale classification problems. To address this problem, we construct a new fast SVM with a truncated squared hinge loss (dubbed as L-SVM). We begin by developing an optimality theory of the nonconvex and nonsmooth L-SVM, which makes it convenient for us to investigate the support vectors and working set of L-SVM. Based on this, we propose a new and effective global convergence algorithm to address the L-SVM. This method is found to enjoy a tremendously low computational complexity, which makes sufficiently decreasing the demand for extremely large-scale computation possible. Numerical comparisons with eight other solvers show that our proposed algorithm achieves excellent performance on large-scale classification problems with regard to shorter computational times, more desirable accuracy levels, fewer support vectors and more robust to outliers.

Global convergence Low computational complexity Optimality theory Support vector machine Support vectors Truncated squared hinge loss

SCI ; EI

英语

其他

National Natural Science Foundation of China[11871183];National Natural Science Foundation of China[11971052];National Natural Science Foundation of China[62106096];National Natural Science Foundation of China[62206120];

Computer Science

Computer Science, Information Systems

WOS:001013595800001

ELSEVIER SCIENCE INC

20232214164989

Computational complexity ; Vectors

Computer Theory, Includes Formal Logic, Automata Theory, Switching Theory, Programming Theory:721.1 ; Computer Software, Data Handling and Applications:723 ; Algebra:921.1

COMPUTER SCIENCE

2-s2.0-85160353917

Scopus

1.Department of Mathematics and Statistics,Changsha University of Science and Technology,Changsha,China
2.School of System Design and Intelligent Manufacturing,Southern University of Science and Technology,Shenzhen,China
3.School of System Design and Intelligent Manufacturing,the Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,China

Wang，Huajun,Li，Genghui,Wang，Zhenkun. Fast SVM classifier for large-scale classification problems[J]. Information Sciences,2023,642.

Wang，Huajun,Li，Genghui,&Wang，Zhenkun.(2023).Fast SVM classifier for large-scale classification problems.Information Sciences,642.

Wang，Huajun,et al."Fast SVM classifier for large-scale classification problems".Information Sciences 642(2023).