Using “Barriers” in Kernel Density Estimation to Improve the Predictive Accuracy of Crime Forecasts: A Case Study of Three Florida Cities
Date
2022-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
MSCC, The University of Tampa
Abstract
Kernel density estimation (KDE) is one of the most popular crime hot spot mapping methods used to reduce and prevent crime. However, this
technique does not consider where crime cannot occur within a study area when a crime risk surface is interpolated. Therefore, a knowledge
gap exists as to how effective incorporating barriers into KDE analysis can be in producing more accurate prospective crime hot spot maps.
Therefore, the current study investigated whether the predictive accuracy of crime forecasts based on KDE will improve when barriers to crime
are incorporated into the analytic process.
Description
item.page.type
Thesis
item.page.format
Keywords
Barriers, Kernel Density Estimation, Predictive Policing, Crime Analysis