Crime Linkage Using Land-use Classification Patterns

• In Event: Poster Session 1

Abstract

Crime linkage using data on patterns of land-use classification is a relatively unexplored area in research. This study aimed to evaluate the similarities in land-use patterns in the neighborhoods where linked crimes (LC) committed by serial offenders. A total of 59 sexual offenders who committed five or more crimes against strangers in the Tokyo metropolitan area were analyzed. An index crime (IC) represented the latest crime in the series, while four preceding crimes with similar characteristics to the IC were considered LCs. In constrast, sexual offences committed by others in proximity to the IC and sharing similar features were identified as unlinked crimes (UC). Two indices were calculated to evaluate similarities between the IC and LCs and between the IC and UCs, using Euclidean distance and land-use classification patterns. Results indicated no significant difference in Euclidean distances from the IC to LCs and UCs. Moreover, patterns of land-use classification in the neighborhoods of the ICs were not similar to those of UCs but rather to those of LCs. We discuss whether patterns of land-use classification contribute to the accurate crime linkage.

Authors

• Kazuki Hirama, National Research Institute of Police Science

• Kaeko Yokota, National Research Institute of Police Science

• Yusuke Otsuka, National Research Institute of Police Science

• Kengo Furuhashi, National Research Institute of Police Science

• Kazumi Watanabe, National Research Institute of Police Science