Downloads: 2
Original Research | Computer Science | Volume 15 Issue 7, July 2026 | Pages: 1684 - 1690 | India
An Efficient DDoS Attack Detection Approach in IoT using a Recurrent Neural Network
Abstract: The rapid growth of the Internet of Things (IoT) has increased the susceptibility of interconnected devices to Distributed Denial-of-Service (DDoS) attacks, necessitating efficient and reliable intrusion detection mechanisms. This study presents a comparative evaluation of three recurrent deep learning (DL) models- Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU)- for DDoS attack detection using the CICIoT2023 benchmark dataset. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied during data preprocessing. Experimental results demonstrate that all three models achieved high detection performance, with the GRU consistently outperforming the RNN and LSTM. The GRU attained 99.22% accuracy, 99.16% precision, 99.28% recall, 99.22% F1-score, and an AUC of 0.9984, indicating excellent discrimination between benign and malicious IoT traffic with a low false alarm rate. Its simplified gating mechanism efficiently captured temporal dependencies while requiring lower computational complexity. These findings demonstrate that the GRU model is an effective and computationally efficient solution for real-time DDoS attack detection in IoT environments.
Keywords: Long Short-Term Memory, Gated Recurrent Neural Network, DDoS Attack Detection, Internet of Things (IoT)
How to Cite?: Dr. S. Anitha, "An Efficient DDoS Attack Detection Approach in IoT using a Recurrent Neural Network", Volume 15 Issue 7, July 2026, International Journal of Science and Research (IJSR), Pages: 1684-1690, https://www.ijsr.net/getabstract.php?paperid=SR26720075346, DOI: https://dx.doi.org/10.21275/SR26720075346