摘要
针对最小二乘支持向量机特征选择及参数优化问题,提出了一种基于PSO的LS-SVM特征选择与参数同步优化算法。首先产生若干种群(特征子集),然后用PSO算法对特征及参数进行优化。在UCI标准数据集上进行的仿真实验表明,该算法可有效地找出合适的特征子集及LS-SVM参数,且与基于遗传算法的最小二乘支持向量机算法(GALS-SVM)和传统的LS-SVM算法相比具有较好的分类效果。
For the feature selection and parameter optimization of LS-SVM,a simultaneous feature selection and LS-SVM parameter optimization algorithm are proposed based on PSO algorithm.At first,a population of particles(feature subsets ) is randomly generated.Then feature and parameter are optimized by PSO algorithms.The experiments on the UCI database are done with this algorithm.Experimental results indicate that this method can efficiently find the suitable feature subsets and LS-SVM parameters.Compared with GALS-SVM and LS-SVM,PSOLS-SVM can get a better classification performance.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第1期134-136,229,共4页
Computer Engineering and Applications
关键词
最小二乘支持向量机
特征选择
参数优化
粒子群算法
Least Squares Support Vector Machines(LS-SVM)
feature selection
parameters optimization
particle swarm optimization algorithm