Hybrid Machine Learning Model for Efficient Malware Network Attack Detection in IoT Environment
Prasad Bobbillapati, S Md Ismail, Syeda Farhath Begum, Farheen Sultana
The exponential growth of the Internet of Things (IoT) has significantly increased the attack surface
for cyber threats, making malware-based network attacks a critical security challenge. Traditional
intrusion detection systems (IDS) often struggle to cope with the high volume, complexity, and evolving
nature of these attacks. To address this, we propose a Hybrid Machine Learning Model that integrates
supervised learning, ensemble techniques, and deep learning-based anomaly detection to enhance the
accuracy and efficiency of malware detection in IoT networks. The proposed model leverages feature
selection, real-time traffic analysis, and hybrid classification to detect malicious network activities while
minimizing false positives. We employ a combination of Decision Tree, Random Forest, and Deep Neural
Networks (DNNs) to classify benign and malicious traffic with high precision. Experimental evaluations
using benchmark datasets demonstrate that our model outperforms traditional IDS models, achieving
superior detection rates, lower latency, and enhanced robustness against sophisticated cyberattacks.
Despite its high efficiency, challenges such as adversarial attacks, scalability concerns, and realtime deployment overhead remain open areas for further research. Future work will explore federated
learning, blockchain-based authentication, and explainable AI (XAI) to further strengthen IoT security.
The proposed hybrid approach provides a scalable, intelligent, and real-time malware detection system,
contributing to a more resilient IoT security framework.
