Enhanced Intrusion Detection in Cloud and IoT Networks Using Motif-Based Machine Learning and Big Data Analytics

This study enhances intrusion detection in cloud and IoT environments by integrating motif-based machine learning with big data analytics. We develop and evaluate a motif-based intrusion detection system (IDS) to improve anomaly detection and inlier classification. Comparative analysis reveals that motif discovery significantly enhances inlier detection, achieving 1.00 precision and 0.95 recall by reducing noise and extracting structured patterns. However, the raw dataset exhibits feature redundancy, leading to lower recall (0.67) and inconsistent anomaly detection. While motif discovery strengthens normal behavior classification, it reduces sensitivity to rare outliers. To mitigate this, we propose integrating targeted outlier detection techniques to balance structured inlier detection with improved anomaly recognition, ensuring a more adaptable and robust IDS for evolving network threats.