Analog circuits diagnosis based on support vector machine with parameter optimization by improved NKCGWO

Abstract

Support Vector Machine (SVM) is a widely used machine learning method in analog circuits fault diagnosis. However, SVM parameters such as kernel parameters and penalty parameter can seriously affect the classification accuracy. The current parameter optimization methods have some defects, such as low convergence speed, easy to fall into local optimal solution and premature convergence. In view of this, an improved grey wolf optimization algorithm (GWO) based on nonlinear control parameter strategy, the first Kepler’ law strategy and chaotic search strategy (NKCGWO) is proposed to overcome the shortcoming of the traditional optimization methods in this paper. In NKCGWO method, three strategies are developed to improve the performance of GWO. Thereafter, the optimal parameters of SVM are obtained using NKCGWO-SVM. To evaluate the performance of NKCGWO-SVM for analog circuits diagnosis, two analog circuits are employed to fault diagnosis. The proposed method is compared with GA-SVM, PSO-SVM and GWO-SVM. The experimental results show that the proposed method has higher diagnosis accuracy than the other compared methods for analog circuits diagnosis.