NCJ Number
212036
Journal
International Journal of Police Science and Management Volume: 7 Issue: 3 Dated: Autumn 2005 Pages: 160-174
Date Published
2005
Length
15 pages
Annotation
This study evaluated the crime-matching ability of three algorithms to deal with patterns of offending which may be a product of local conditions and the value of modus operandi (MO) information as an effective means of matching a target crime to the offender, specifically burglary.
Abstract
Research has long recognized the value of modus operandi (MO) information in the investigation of violent and sexual crimes. MO information is always present in some degree at a crime scene. Utilizing the potential of MO information in the detection of burglary is important as this volume crime has a low detection rate despite its concentration in terms of location and large percentage of offenses committed by a small proportion of offenders. This study evaluated the crime-matching ability of 3 algorithms to match a target crime to the actual offender within a database of 966 offenses committed by 306 offenders. The three algorithm models are: (1) a random choice probability using behavioral information (RCPA); (2) an algorithm based on recency, prolificness, and actual location information (RPAL); and (3) an algorithm that is a combination of the other two (COMBIN). The algorithms were evaluated based on behavior and geographic/temporal information. The results clearly demonstrate the value and means of utilizing such data in linking crimes. Graphs, tables, and references