A GWO based efficient approach to identify terrorist incident hotspots in India

For a given set of geospatial locations (like, crime activities, terrorist activities, bomb blast locations, etc.), identification of such circular zones where accumulation of points inside the circle is very much greater than outside is important. Such zones are known as hotspots and their detection is known as circular hotspot detection (CHD). Timely detection of circular hotspots is crucial in many societal applications like epidemiology, terrorism, criminology etc. The state-of-the-art method for circular hotspot detection viz. SaTScan is computationally expensive due to enumeration of all possible circles called candidate circular hotspots. Due to its high cost SaTScan is not suitable for applications like terrorist activity hotspot identification, where well-timed identification of hotspots is crucial to prioritize the security efforts put by government and security agencies. Therefore, in this paper, we present an efficient and effective Grey Wolf Optimizer based approach called GWO-CHD for terrorism hotspot detection. The results of GWO-CHD are compared with SaTScan in terms of time required to detect the hotspot and its quality (measured using relative error). All the experiments are performed using terrorist activity data of Indian subcontinent from 2016-2021. Results indicate that hotspots identified by GWO-CHD and SaTScan are almost at par in terms of quality; however, GWO-CHD proved to be much more efficient than SaTScan in terms of computational time.