A random forest framework for automated fault classification in analog filter circuits using simulation-derived features

The identification of faults in analog filter electronic systems has been challenging due to their non-linear characteristics, sensitivity to parameter variations and complex component interactions. Traditional fault diagnosis methods often involve lengthy testing procedures and fail to provide reliable results for modern complex electronic systems. In this work, an automated fault classification framework based on machine learning is proposed for analog filter circuits is relatively new. Circuits considered are filters like Sallen–Key and Chebyshev .The circuits are simulated under both normal and faulty conditions and key features are extracted from their responses, such as natural frequency ( ω n ), damping ratio ( ζ ), gain ( G 0 ). A two-stage Random Forest-based machine learning model is employed, where the first stage identifies the circuit type and the second stage performs fault classification specific to that circuit. The proposed method achieves an overall accuracy of approximately 98.5% for both fault component and fault type classification.