Parameter Optimization for SVM Classification Based on NGA

Qin, Hanshu; Yan, Wei; Zeng, Shaohua

doi:10.1007/978-1-4471-4856-2_69

“…Zhou et al [14] put forward a new genetic algorithm with improved genetic operators (IO-GA) to optimize the SVM classifier's parameters. Qin et al [15] formulated parameter optimization of SVM for classification based on niche genetic algorithm (NGA) which avoided the premature situation and better maintained the diversity of solution.…”

Section: Genetic Algorithm (Ga)mentioning

confidence: 99%

A Brief Overview on Parameter Optimization of Support Vector Machine

Cao¹,

Yu²,

Cheng³

2017

dtmse

View full text Add to dashboard Cite

Support vector machine (SVM) has been successfully applied in classification and regression problems. But it is very sensitive to the selection of parameters. The fundamental principles of SVM are analyzed firstly. The main optimization methods and achievements for SVM parameters are introduced. And the popular fitness functions used for the parameter optimization of SVM are described. The objective of this paper is to provide readers a brief overview of the recent advances for parameter optimization of SVM and enable them to develop and implement new optimization strategies for SVM-related research at their disposal.

show abstract

“…ii) Qin et al [15] evaluated the classification performance of SVM with new leave-one-out known as NLOO. The fitness function is shown as below:…”

Section: Fitness Functions For Parameter Optimization Of Svmmentioning

confidence: 99%

A brief overview on parameter optimization of support vector machine

Cao

¹

,

Yu

²

,

Cheng

³

2014

Computational Intelligence and Industrial Engineering

0

View full text Add to dashboard Cite

Support vector machine (SVM) has been successfully applied in classification and regression problems. But it is very sensitive to the selection of parameters. The fundamental principles of SVM are analyzed firstly. The main optimization methods and achievements for SVM parameters are introduced. And the popular fitness functions used for the parameter optimization of SVM are described. The objective of this paper is to provide readers a brief overview of the recent advances for parameter optimization of SVM and enable them to develop and implement new optimization strategies for SVM-related research at their disposal.

show abstract