Kernel density estimation and hotspot mapping: Examining the influence of interpolation method, grid cell size, and bandwidth on crime forecasting

The purpose of this paper is to examine the effects of user-defined parameters settings (e.g. interpolation method, grid cell size, and bandwidth) on the predictive accuracy of crime hotspot maps produced from kernel density estimation (KDE).

The influence of variations in parameter settings on prospective KDE maps is examined across two types of interpersonal violence (e.g. aggravated assault and robbery) and two types of property crime (e.g. commercial burglary and motor vehicle theft).

Results show that interpolation method has a considerable effect on predictive accuracy, grid cell size has little to no effect, and bandwidth as some effect.

The current study advances the knowledge and understanding of prospective hotspot crime mapping as it answers the calls by Chainey et al. (2008) and others to further investigate the methods used to predict crime.