Fault Diagnosis of Circuits with Tolerance Based on Support Vector Machines

Abstract

There are many difficulties in diagnosis of analog circuits with tolerance because of the uncertainty characteristic the circuits, it is proposed to construct fault classifiers using the support vector machines (SVMs) algorithm. The fault classifiers based on SVMs can realize precise fault diagnosis even when a few samples are gotten, they have better generality and practicality too. Experiment shows the diagnosis method based on SVMs technique for analog circuits with tolerance types can completely overcome the limitations from some conventional classification methods such as neural networks, achieve nonlinear partition and solve the essential recognition problem for analog circuits with tolerance fault types