Analog Circuit Intelligent Fault Diagnosis Based on Greedy Kpca and One-against-rest Svm Approach

Fault diagnosis of analog circuits is essential for guaranteeing the reliability and maintainability of electronic systems. A novel analog circuit fault diagnosis approach based on greedy kernel principal component analysis (KPCA) and one-against-rest support vector machine (OARSVM) is proposed in this paper. In order to obtain a successful fault classifier, eliminating noise and extracting fault features are very important. Due to better performance of nonlinear fault features extraction and noise elimination, KPCA is adopted as a processor. However, a drawback of KPCA is that the storage required for the kernel matrix grows quadratically, and the computational cost for eigenvector of the kernel matrix grows linearly with the number of training samples. Therefore, greedy KPCA which can approximate KPCA with small representation error is introduced to enhance computational efficiency. Based on statistical learning theory and the empirical risk minimization principle, SVM has advantages of better classification accuracy and generalization performance. The extracted fault features are then used as the inputs of OARSVM to solve fault diagnosis problem. The effectiveness of the proposed approach is verified by the experimental results.