Significant research has been conducted in predictive policing, with the intention of reducing crime. However, many studies focus on the accuracy of prediction, without recognizing that the utility of a predictive model depends on the factors upon which the model is built.
If the factors are not controllable, then the model cannot be directly used to take action. Further, the policing system is highly dynamic over space and time. If a police department employs an annual staffing allocation model, then the dynamic deployment strategy required to continually react to crime predictions may not be feasible.
The goal of the proposed research is to build a framework to bridge this gap between predictive policing and dynamic strategies. The purpose is to create a discrete-event simulation tool to facilitate assessment of dynamic strategies. The research questions are: (1) What are the key structural components of dynamic strategies? (2) What constraints limit them? (3) How should predictive policing be incorporated in our simulation tool? (4) What additional data are needed to inform dynamic strategies?
The researcher has been collaborating with the Arlington, Texas Police Department, which has provided motor vehicle burglary case data and staffing data for the proposed framework. Additional data may be obtained, if available.
This project contains a research and/or development component, as defined in the applicable law.
nca/ncf
