Application of IWO-SVM approach in fault diagnosis of analog circuits

Abstract

Support vector machine (SVM) is a machine learning algorithm which has been applied to fault diagnosis of analog circuits. Invasive weed optimization (IWO) is a novel numerical optimization algorithm inspired from weed colonization. An approach that combines IWO and SVM (IWO-SVM) is proposed to fault diagnosis of analog circuits in this paper. The process of fault diagnosis of analog circuits using IWO-SVM approach is introduced in details. A biquadrate filter is used to test the performance of IWO-SVM approach for fault diagnosis. The simulation experiments show that the IWO-SVM approach proposed in this paper has a higher diagnosis accuracy rate than the conventional SVM in fault diagnosis of analog circuits.