An Efficient Approach Using Lightweight Machine Learning Models for Detection of DDoS Attacks on IoT Devices

Advancements in the Internet of Things (IoT) have introduced significant security challenges, with Distributed Denial of Service (DDoS) attacks emerging as one of the most critical threats. These attacks involve botnets controlled by attackers that flood networks with malicious traffic, disrupting legitimate services. As the global DDoS landscape evolves, it has become ever more critical for IoT devices to rapidly identify and mitigate these threats. Current machine learning-based detection methods are often too complex for IoT devices with limited resources. Existing solutions usually rely on overly complex models, overlooking the opportunity to use lightweight and efficient techniques for real-world deployment. To address these limitations, this research proposes a novel and efficient framework that combines feature selection with lightweight machine learning models, including Random Forest, Logistic Regression, and Naive Bayes, for effective DDoS detection in IoT environments. Evaluated on the NSL-KDD dataset, the results demonstrate that the proposed framework significantly outperforms existing methods, achieving 99.88% detection accuracy with the Random Forest model, 91.61% with Logistic Regression, and 87.62% with Naive Bayes. This research advances IoT security by integrating feature selection with lightweight machine learning, providing practical and effective solutions for wireless applications.