A Supervised Intrusion Detection System Leveraging Machine Learning for Secure Smart Education in the Internet of Education (IoEd)

Le résumé fourni par la source

The Internet of Education (IoEd) is an emerging field of IoT that combines Internet of Things (IoT) technologies with various learning environments such as smart classrooms, LMS, and web applications to improve educational productivity more efficiently and effectively. However, the integration of IoT Technologies within the educational environment, led to significant security risks, like different cyber-attacks, especially due to their limited processing capacity, which makes traditional defenses like firewalls and VPNs less effective. A single vulnerability within an IoT device can compromise the integrity of the entire network. Thus, this study highlights the urgent need to secure IoT-enabled academic systems to support sustainable educational advancement. This study proposes an Intrusion Detection System (IDS) based on Machine Learning, that identifies abnormal activities and potential cyber threats in an academic network, using a recently curated dataset, IoEd-Net. Leveraging Python packages like scikit-learn, TensorFlow, pandas, and NumPy. Six ML classifiers are used; Logistic Regression, Decision Tree, AdaBoost, Random Forest, ANN, and KNN. The experiment was done into two ways, 70/30 train-test split had achieved the highest accuracy of 98\% for Random Forest. The same highest accuracy was obtained in a second run with 5-fold cross-validation, by contributing to the IoEd security by developing an Intrusion Detection system (IDS) that can predict unseen network activities. It gives valuable analysis into network behavior, also enhancing privacy and security in an IoEd networks.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.