特征选择与支持向量机参数同步优化研究

Synchro-Optimization of Feature Selection and Parameters of Support Vector Machine

提出了免疫多向二进制粒子群优化算法。基于该算法实现了特征选择与支持向量机参数的同步优化,克服了单独优化特征或单独优化支持向量机参数的缺陷。既解决了特征与分类器不匹配带来的诊断能力下降,又提高了故障诊断精度与搜索速度。

An algorithm named immune multi-direction binary particle swarm optimization (IMBPSO) algorithm was presented and applied to optimize feature selection and parameters of support vector machine(SVM) simultaneously.It overcomes the degression of diagnosis ability resulting from unmatch of the features and the classifier parameters and improves the diagnosis precision and search speed.An example of engine fault classification demonstrates the effectiveness the method.