地理的プロファイリング支援ソフトウェアの開発-被疑者拠点空間分析システム SAMON -

Abstract

Existing geographic profiling software that performs the widely tested probability distance strategies has issues when implemented in criminal investigation in Japan. Therefore, we developed the Spatial Analysis Methods of Offender’s Nodes (SAMON) software based on a free software environment, R. Given the issues involving existing software, SAMON includes the following three features: (1) prediction of an offender’s home base using different distance decay functions constructed from Japanese burglars’ Journey-to-Crime distances; (2) validation of prediction accuracy in the solved case; and (3) calibration of the distance decay functions using a sample of solved cases in a type and region that the user is interested in. We expect that SAMON will improve the availability of probability distance strategies and its accuracy in the Japanese context.