AI-Driven Cyber Threat Prediction System Using Dynamic Graph-Based Anomaly Detection in Enterprise Networks
Sourav Mandal
Enterprise networks are increasingly exposed to dynamic cyber threats capable of bypassing traditional intrusion detection systems. This research proposes a novel AI-driven threat prediction framework using dynamic graph-based anomaly detection. The model represents enterprise network communication as a continuously evolving graph and applies temporal graph learning, unsupervised anomaly scoring, and behavior modeling to predict malicious activities in real time. The system does not rely on prior attack signatures and adapts automatically to new network behavior. Experimental results demonstrate improved detection accuracy, reduced false positives, and enhanced robustness against zero-day attacks, proving the effectiveness of dynamic graph learning for next-generation enterprise cybersecurity.