A Convolutional Neural Network-Based Deep Learning Framework for Automated Cyberattack Detection in IoT Applications

Sivananda Hanumanthu; G. Anil Kumar

doi:10.62647/IJITCE2025V13I2PP1453-1461

Authors

Sivananda Hanumanthu Research Scholar, Department of CSE, Bharatiya Engineering Science and Technology Innovation University (BESTIU), Andhra Pradesh & Director of Enterprise Architecture at Rubrik, Bangalore, India Author
G. Anil Kumar Principal & Professor Dept of CSE, Scient Institute of Technology, Hyderabad, TS, India. Author

DOI:

https://doi.org/10.62647/IJITCE2025V13I2PP1453-1461

Keywords:

Network Security, IoT Security, Artificial Intelligence, CNN-Based Threat Detection, Intrusion Detection System

Abstract

With the combination of diverse IoT technologies
and no global standards imposed, the IoT has further
opened up revolutionary, innovative applications yet
has also presented new and complex security
challenges. Restrictions are needed to safeguard IoT
applications since cyber threats are increasing daily.
Artificial Intelligence (AI) has made it possible to
solve many real-world issues. This study introduces
the Learning-based Cyberattack Detection system
(LbCADF). This deep learning-based system
employs a CNN-based model with enhanced
sensitivity for autonomously detecting cyberattacks
in IoT settings. The framework successfully
distinguishes between benign and malevolent traffic
flows. We add feature selection and hyperparameter
tweaking to improve training quality to our proposed
system, Enhanced CNN for Attack Detection and
Classification (ECNN-ADC). To prevent
overfitting, an early stopping criterion is applied.
This work has been evaluated on the benchmark
dataset UNSW-NB15. At the same time, the
empirical results indicate that ECNN-ADC achieves
the highest detection accuracy (95%) compared to
various of the latest models (MLP, baseline CNN).