Automatic Parametric Fault Detection in Complex Microwave Filter Using SVM and PCA

The aim of this paper is to present the diagnostics of complex linear analog systems with parametric faults, using Support Vector Machine (SVM) as a tool for fault location. The diagnostic results of a microwave filter with the help of SVM network are presented. A strategy for finding the optimal kernels and their parameters for the particular system under test is proposed. A method for characteristic points reduction based on the statistical PCA method is also presented.