In order to accurately track the changing characteristics of analog circuit fault,an analog circuit fault diagnosis model based on feature and weighting selection of artificial fish swarm algorithm(AFSA)is put forward. The original feature set of the analog circuit state is obtained according to the Volterra series. The relevant vector machine is adopted as the classifier of the analog circuit fault. The artificial fish swarm algorithm is used to select the important feature subset,and give a rational weight for each feature. The model was applied to a certain analog fault circuit. The results show that the artificial fish swarm al⁃gorithm can get the optimal feature subset accurately,the analog circuit fault rate is averagely higher than 95%,and the perfor⁃mance of the model is significantly superior to the classical model.%为了准确跟踪模拟电路故障的变化特点，提出一种人工鱼群算法选择特征和加权的模拟电路故障诊断模型。首先根据Volterra级数获得模拟电路状态的原始特征集，然后采用相关向量机作为模拟电路故障的分类器，采用人工鱼群算法选择重要特征子集，并赋予每一个特征合理权值，最后将该模型应用于某一模拟电路故障中。结果表明，人工鱼群算法可以准确得到最优特征子集，模拟电路故障平均超过95%，而且其性能要显著优于经典模型。
展开▼
