Prospective spatio-temporal data analysis for security informatics

Spatio-temporal data analysis plays a central role in many security-related applications including those relevant to transportation infrastructure and border security. In this paper, we investigate prospective spatio-temporal analysis methods that aim to identify "unusual" clusters of events, or hotspots, in both spatial and temporal dimensions. We propose a Support Vector Machine-based approach and compare it with a well-known prospective method based on space-time scan statistic using three problem scenarios. The first two scenarios are based on simulated data with known hotspots. The third scenario uses a real-world crime analysis data set involving vehicles.