Breaking the trend: Anomaly detection models for early warning of socio-political unrest

This paper presents an innovative Early Warning System for predicting conflicts and unrest based on Anomaly Detection, identifying sudden and unexpected changes in behavioral patterns that may indicate the potential for these events to occur. This approach draws inspiration from various fields – including industry, such as manufacturing, physics and networking – but its application in the domain of diplomacy is entirely new. The system, tested on three case studies, showcase its ability to enhance open-source intelligence technique in the diplomatic arena. The study provides a fresh perspective on predictive analytics and focuses on examining outbreaks.