摘要
为解决支持向量机算法(Support Vector Machine,SVM)的核函数参数及惩罚因子参数选取的盲目性,利用果蝇优化算法(Fruit Fly Optimization Algorithm,FOA)对SVM中参数进行优化。提出基于FOA的SVM故障诊断算法,并对汽轮机故障实验数据进行模式识别。该算法能对SVM相关参数自动寻优,且能达到较理想的全局最优解。通过与常用的粒子群算法(Particle Swarm Optimization,PSO)与遗传算法(Genetic Algorithm,GA)优化后支持向量机进行对比。结果表明,FOA-SVM算法稳定、识别速度快、识别率高。
In order to solve the problem that the selection of the kernel function parameters and penalty factor parameters in the support vector machine（SVM）algorithm is blindfold,the fruit fly optimization algorithm （FOA）was applied to optimize the parameters in SVM.A fault diagnosis algorithm of SVM based on FOA was put forward,and then the pattern recognition of experimental turbine failure data was performed.The algorithm can optimize the SVM parameters automatically,and achieve ideal global optimal solution.Comparing with the SVMwhich is optimized by the commonly used methods of the particle swarm optimization（PSO）and the Genetic Algorithm （GA）,the results demonstrate that FOA-SVM has the fastest recognition speed and the highest recognition rate.
出处
《振动与冲击》
EI
CSCD
北大核心
2014年第22期111-114,共4页
Journal of Vibration and Shock
基金
吉林省科技发展计划项目(20100506)
关键词
支持向量机
汽轮机
振动诊断
果蝇算法
support vector machine
turbine
vibration diagnosis
fruit fly optimization algorithm
