Hybrid deep CNN RNN model for securing IOT against DDOS attacks
Ameer Ali Jawad, Ahmad Ghandour, Asaad Abdul Malik Madhloom Al-Salih, Mohammed Abdul Majeed, Mahmood Anees Ahmed, Bashar Ahmed Khalaf, Ibraheem Kasim Ibraheem, Ahmad Taher Azar, Amjad J. Humaidi, Mohammed Majid Msallam · Pollack Periodica · 2025
Abstract The rapid expansion of Internet of things networks has increased vulnerability to distributed denial of service attacks. This paper proposes a hybrid deep learning model that combines convolutional neural networks and recurrent neural networks for effective distributed denial of service detection in Internet of things environments. The model leverages convolutional neural networks for feature extraction and recurrent neural networks for temporal modeling to classify benign, light, and heavy attack traffic. Evaluated using the CIC-Bell-DNS-EXF2021 dataset, it achieved 99.5% accuracy, 99.9% precision, and 99.6% F1 score, outperforming traditional machine learning methods and enhancing real-time security.