AI driven cyber threat classification and risk assessment for IoT ecosystem security

The rapid expansion of Internet of Things (IoT) ecosystems has increased exposure to cyberattacks, requiring real-time intrusion detection, vulnerability analysis, and automated risk evaluation. The proposed security framework combines behavioral anomalies detected by a cloud-based machine learning-driven Network Intrusion Detection System (ML-NIDS), structural weaknesses identified through Automated Vulnerability Scanning (AVS), and standardized Common Vulnerability Scoring System (CVSS) severity metrics to enable continuous and scalable threat monitoring for IoT networks. Network traffic and vulnerability scan results generated at the edge using a Raspberry Pi are securely transmitted to Amazon Simple Storage Service (Amazon S3) and analyzed using a Random Forest machine learning classifier deployed on Amazon Web Services (AWS) SageMaker. The proposed model achieves up to 99% detection accuracy for major attack categories, including Distributed Denial of Service attacks, spoofing, and Mirai botnet activity. A unified risk scoring engine correlates machine learning-based anomaly scores, AVS outputs, and CVSS severity metrics to quantify real-time threat impact. Experimental results demonstrate that the proposed architecture achieves high detection accuracy and provides scalable, context-aware risk assessment for dynamic IoT environments. By integrating edge-based monitoring with cloud-driven analytics and automated vulnerability intelligence, the framework significantly strengthens cyber resilience in IoT deployments.