<![CDATA[Cyberattack detection in Electric Vehicle Charging Infrastructure (EVCI) is increasingly critical as the global transition toward electric mobility accelerates to reduce carbon emissions. This study provides a comprehensive evaluation of machine learning models for cyberattack detection using the CICSEV2024 dataset. The performance of tree-based algorithms, including Decision Trees (DT), Random Forest (RF), and Gradient Boosting (GB), is compared to identify effective yet interpretable models. Experimental results demonstrate that these models achieve exceptional performance, with DT, RF, and GB reaching 100% accuracy and precision. Furthermore, 10-fold cross-validation on an imbalanced dataset (Benign class) confirms the models’ consistency, maintaining a score of 1.00 across all iterations. The proposed models also achieve a perfect Area Under the Curve (AUC) score of 1.00, indicating their robustness and reliability in detecting cyberattacks. The findings highlight that simple and interpretable tree-based models can achieve state-of-the-art performance in EVCI cybersecurity detection, offering practical implications for enhancing the security of electric vehicle charging infrastructures in real-world deployments.]]>
Copyrights © 2025
