Interval-Censoring Methods for Aoristic Crime Analysis

• Activity Number: 470
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Defense and National Security
• Author(s): Michael D. Porter*+
• Companies: University of Alabama
• Keywords: crime ; interval censoring ; point process ; aoristic ; smoothing ; self-exciting

Abstract

For certain types of crime (e.g., burglary, auto theft) the exact timing of a crime event is often unknown because the victim was not present when the crime was committed. In these situations, the police usually record the time interval in which the crime could have occurred. Aoristic analysis is a common approach to estimate the temporal crime rate in a region when the exact timing of the individual crime events is subject to this type of uncertainty. However, the aoristic approach can oversmooth the rate and conceal interesting patterns. To alleviate this bias and improve estimation, we present a new method based on interval censoring techniques. By incorporating information from the other crime events, our approach goes a step in the direction of modeling the offender behavior rather than the routine activities of the victims.