基于选择性SVM集成的模拟电路故障诊断方法

摘要

为克服支持向量机在故障诊断应用中存在的不足,设计了基于聚类分析的选择性支持向量机集成学习算法,并应用于模拟电路故障诊断.该方法采用K-means聚类算法去除相似冗余个体,提高剩余个体学习机的差异性,增强了支持向量机集成模型的泛化能力.以ITC' 97标准电路中的Leap-Frog滤波电路为诊断实例进行了仿真实验.%A method of analog circuit fault diagnosis based on selective SVM ensemble is pres-ented in this paper. K - means clustering algorithm is used to improve the diversity of individuals in SVM ensemble, he method overcomes disadvantages of single SVM and greatly improves the generation ability. Simulation experiments on a Leap -Frog filter circuit are carried out.