基于贝叶斯证据框架下的最小二乘小波支持向量机,设计了一种新型模拟电路故障诊断方法。将贝叶斯证据框架应用于多类LS-WSVM分类器来选取正规化参数和核参数,并采用小波提升变换对从测试点得到的各种故障状态下输出电压信号进行分解获取多尺度的小波系数,对经处理的小波系数提取出故障特征量,以此作为样本训练多类LS-WSVM分类器来确定模拟电路故障诊断的模型。采用雷达系统模拟电路进行了仿真,结果表明,设计的模拟电路的故障诊断方法效果良好。 Based on the least square wavelet support vector machine under Bayesian evidence framework, a novel fault diagnosis method for analog circuits is designed. The Bayesian framework is applied to many kinds of LS-WSVM classifiers to select regularization parameters and kernel parameters. The wavelet transform is used to decompose the output voltage signals under various fault states obtained from the test points to obtain multi-scale wavelet Coefficient, the extracted wavelet feature of the fault feature extraction, as a sample training multiple LS-WSVM classifier to determine the analog circuit fault diagnosis model. The simulation of the radar system was carried out. The results show that the fault diagnosis method of the analog circuit is effective.