Anomaly Detection in IoT Networks Using Machine Learning Techniques

Abstract

The rapid growth of IoT introduces significant security challenges, necessitating effective anomaly detection techniques. This paper implements a Random Forest Classifier for detecting and classifying anomalies in IoT network traffic using the RT_IOT2022 dataset. The model achieves 99.8% accuracy with high precision, recall, and F1-scores across multiple attack types (e.g., MQTT_Publish, DDoS). Detailed evaluation confirms the classifier's effectiveness in distinguishing diverse attacks, demonstrating the viability of machine learning for enhancing IoT security. This work contributes to developing resilient IoT systems, though future research should address class imbalance and comparative performance with other models