摘要
支持向量机(SVM)是统计学习理论(SLT)的一种成功实现,它建立在SLT的VC维理论和结构风险最小化原理基础之上,能够较好地克服神经网络容易出现的过学习和泛化能力低等缺陷.提出了一种基于支持向量机的装备故障诊断方法,应用该方法成功地对某型装备几种典型故障进行了正确诊断.在对检验样本施加噪声后,支持向量机构成的故障分类器仍然能够满足发动机故障诊断要求,表明提出的故障诊断算法具有良好的鲁棒性和推广性.
Support vector machine is realized in statistical learuing theory （SLT） successfully which is established on VC dimension of SLT and structure risk minimization, and can overcome the defaults of over- fitting and lower of generalization ability for neural network. It is shown that a method＇of fault diagnosis for equipment based on support vector machine can diagnose some classical fault for the equipment. The support vector machine can also meet fault diagnosis requirement when noise is applied. It is shown that the methods have a good robustness and popularization.
出处
《战术导弹技术》
2009年第6期26-29,57,共5页
Tactical Missile Technology
关键词
支持向量机
统计学习理论
故障诊断
鲁棒性
support vector machine
statistical learning theory
fault diagnosis
robustness
