https://doi.org/10.1051/epjconf/202532801032
Hybrid Deep Learning-Based Security Model for Robust Intrusion Detection in IoT Networks
1 Research Scholar, School of Computer Science & Engineering, Sandip University, Nasik, Maharashtra, India
2 Assistant Professor, School of Computer Science & Engineering, Sandip University, Nasik, Maharashtra, India
* Corresponding author: Jayashribhoj@gmail.com
Published online: 18 June 2025
The popularity of Internet of Things (IoT) devices has been responsible for a major growth in cybersecurity risks across sectors. This increasing complexity emphasizes the immediate need for more versatile and advanced intrusion detection systems. Our study defines a Hybrid Deep Learning-Based Security Model (HDLSM) meant to solve such problems by effectively distinguishing between possibly malicious and benign IoT network traffic using Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNN). Training and validation of the model was done using the IoT23 dataset, which is a thorough set of real-world, labeled network data covering various malware attacks, including Mirai, Gafgyt, Tsunami, and Torii. To ensure the inputs were of the best quality, we conducted a thorough preprocessing stage including data cleaning, format standardization, and simplification of complex attributes. As we tested the HDLSM model, it achieved 96.6% accuracy, 96.6% precision, 96.1% recall, 96.3% F1 score, and 97.1% AUCROC.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
