Enhancing IoT Security Through Advanced Machine Learning Models for Anomaly Detection

The rapid proliferation of the Internet of Things (IoT) has brought unprecedented connectivity, transforming industries and daily life. However, this integration comes with critical cybersecurity challenges, exposing IoT networks to sophisticated threats like Distributed Denial-of-Service (DDoS) attacks and other malicious activities. This paper investigates IoT vulnerabilities by analyzing the CICIoT2023 dataset and employing advanced machine learning techniques, specifically Random Forest and XGBoost, to detect and mitigate anomalies in network traffic. Through rigorous preprocessing, feature selection, and model tuning, both algorithms demonstrated exceptional performance, achieving accuracy and F1 scores exceeding 99%. XGBoost outperformed Random Forest in key metrics, showcasing its scalability and precision for complex datasets, while Random Forest proved advantageous for resource-constrained environments due to its simplicity and interpretability. This research not only highlights the strengths of machine learning in IoT security but also provides actionable insights for real-time anomaly detection and adaptive defenses. Future work aims to enhance these models for real-world deployment, adaptive learning, and broader applicability across diverse IoT ecosystems, ensuring a secure and resilient interconnected landscape.