A Clustering-Based Method for Hotspot Recognition in Crime Forecasting

Researchers have discovered that crime exhibits regional tendencies since the area of crime prediction has emerged. These patterns enable authorities to take preventative action by helping to anticipate crimes before they occur. Applications for crime prediction include hotspot assessment, policing prediction, and spatial profiling. Every type of crime has some relationship to time, place, weather, and census factors including the area's literacy rate and annual income. These all function as predictors of criminal activity. Crime is predicted in this research by using historical crime episodes as indicators. This research presents a 2-Dimensional Hotspot evaluation-based machine learning-driven spatiotemporal crime prediction technique. To find 2-Dimensional Hotspots, clustering is employed. The proposed technique's performance is assessed using sophisticated classification techniques both with and without analysing hotspot, and it seems that the model including analysis of hotspot performs better.