Enhancing the Performance of Support Vector Classifier Through Hyperparameter Optimization for Electrical Fault Classification

Electrical fault detection and classification is crucial for the safety, stability and reliability of the power system. Accurate classification helps to identify fault types, enabling faster restoration of service. In this research work, a model is proposed based on support vector classification (SVC), with significantly higher evaluation metrics compared to other existing model for transmission line fault classification. After optimizing the hyperparameters of support vector classification (SVC), a 99.3% accuracy rate was achieved. A cross-comparison of the evaluation metrics for different algorithm has been presented to validate the selection of SVC over other algorithms. Also, the evaluation metrics of SVC, both before and after hyperparameter optimization, are compared and analyzed to justify the effectiveness of hyperparameter tuning in SVC.