Back to Departmental Colloquium: Spring 2011
Departmental Colloquium
Date: Thursday, Feb 17, 2011
Time: 4:15PM
Location: JWB 335
Martin Short
University of California, Los Angeles
Title |
The Mathematics of Crime Hotspots |
Abstract |
Criminologists have often observed that crimes tend to form “epidemic” like spatio-temporal patterns, whereby criminal events seem to spawn further events in nearby space-time regions. This often leads to the formation of crime hotspots, which are localized regions of intense criminal activity. However, descriptive and/or predictive mathematical models of these hotspots have been lacking. In this talk, I will describe two possible models for this phenomenon. One starts from the bottom up, modeling some basic aspects of human criminal behavior and arriving at a set of PDEs reminiscent of others in the mathematical biology literature. I will discuss some of the results and implications of the analysis of these PDEs, especially in terms of police response to hotspot formation. The second model starts from the top down, taking a data-driven statistical approach to the problem. This method is closely related to others used in the modeling and prediction of earthquakes, and yields practical results that can help police to best utilize their limited resources. |
Mathematical Biology
Differential Equations
Data Science and Machine Learning