Learning to rank spatio-temporal event hotspots

Crime, traffic accidents, terrorist attacks, and other space-time random events are unevenly distributed in space and time. In the case of crime, predictive policing algorithms aim to focus limited resources at the highest risk crime hotspots in a city. A crucial step in the implementation of these strategies is the construction of scoring models used to rank spatial hotspots. While these methods are evaluated by area normalized Recall@k (called the Predictive Accuracy Index), models are typically trained via maximum likelihood or rules of thumb that may not prioritize model accuracy in the top k hotspots. Furthermore, current algorithms are defined on fixed grids that fail to capture risk patterns occurring in neighborhoods and on road networks with complex geometries. We introduce CrimeRank, a learning to rank boosting algorithm for deterniming a crime hotspot map that directly optimizes the percentage of crime captured by the top ranked hotspots. The method also employs a floating grid combined with a greedy hotspot selection algorithm for accurately capturing spatial risk in complex geometries. We illustrate the performance using crime and traffic incident data provided by the Indianapolis Metropolitan Police Deparment and IED attacks in Iraq.