Comparative Analysis Of Two Variants Of The Knox Test: Inferences From Space-Time Crime Pattern Analysis

This paper compares two variants of the Knox test in relation to space-time crime pattern analysis. A case study of burglary and 'stolen-vehicle' crime data sets of San Francisco city is presented. The comparative analysis shows that while one variant is designed to detect the sizes of the spatio-temporal neighbourhoods at which clustering (hotspots) is prominent within a data set, the other variant is able to reveal the spatial and temporal windows/bands at which crime events are frequently repeated to form clusters (hotspots) across an area.