The precision, accuracy and efficiency of geographic profiling predictions:a simple heuristic versus mathematical algorithms

Bennell, Craig and Emeno, K and Snook, Brent and Taylor, Paul and Goodwill, Alasdair (2009) The precision, accuracy and efficiency of geographic profiling predictions:a simple heuristic versus mathematical algorithms. Crime Mapping: A Journal of Research and Practice, 1 (2). pp. 65-84. ISSN 2152-9876

[img]
Preview
PDF
2010_CM_Bennell_et_al..pdf - Published Version

Download (675kB)

Abstract

This study compared the precision, accuracy, and efficiency of geographic profiles made by students to those made by mathematical algorithms. After making predictions on 20 maps, each depicting a different offense series, nearly half of the sampled students were instructed that "the majority of offenders commit offenses close to home". All of the students were then asked to make predictions on a different set of 20 maps. Seven different mathematical algorithms, several derived from a new Bayesian journey-to-crime estimation method, were also applied to the 40 maps. Results showed that informing students about the "distance decay heuristic" increased the precision of their predictions, but these predictions were not as accurate or efficient as those made by most of the algorithmic procedures. Implications of these results for the field of geographic profiling are discussed.

Item Type:
Journal Article
Journal or Publication Title:
Crime Mapping: A Journal of Research and Practice
ID Code:
49927
Deposited By:
Deposited On:
21 Sep 2011 08:01
Refereed?:
Yes
Published?:
Published
Last Modified:
02 Jul 2020 00:41