Method of Analog Circuit Diagnosis Based on ECOC and SVM

In order to reduce fault diagnosis time for analog circuit based on SVM method and improve the accuracy of fault diagnosis,this paper proposes a method of analog circuits diagnosis by combining error-correcting output code(ECOC) and support vector machine(SVM).Firstly,Fuzzy C-means(FCM) algorithm is employed to cluster the fault samples,and then ECOC matrix is obtained from binary tree.Secondly,the multi-class fault classifiers are designed using SVMs to train and test samples corresponding to ECOC matrix.Finally,test results are decoded by a decoding method.The experimental results show that the diagnosis accuracy of the proposed fault classifier is superior to that of the conventional SVM and BP neural network,and the diagnosis time of the new method is lower than that of the multi-class method based on SVM.