Anomaly Detection in IoT Networks using Graph Neural Networks
M. Gayathri, Vanapalli Veera Snigdha, Yendreddy Jayadurga · 2025
The importance of anomaly detection in IoT device networks has gained much significance in enhancing security, enabling reliable monitoring and reducing potential threats. A deep learning model for detecting different types of cyberattacks on various devices using the autoencoder model and distinguishing normal patterns of traffic from malicious ones by using the RT-IoT2022 dataset is proposed. The proposed scheme uses GNN to perceive the environment for the identification and classification of real-time cyberattacks. It makes IoT devices more secure, stable and achieves high accuracy.The proposed GNN-based algorithm demonstrated an impressive 99% accuracy in the identification of anomalies in IoT networks, providing strong security and threat avoidance.