| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 98439 | Forensic Science International | 2007 | 5 Pages |
Abstract
Pattern recognition techniques can be very useful in forensic sciences to point out to relevant sets of events and potentially encourage an intelligence-led style of policing. In this study, these techniques have been applied to categorical data corresponding to cutting agents found in heroin seizures. An application of graph theoretic methods has been performed, in order to highlight the possible relationships between the location of seizures and co-occurrences of particular heroin cutting agents. An analysis of the co-occurrences to establish several main combinations has been done. Results illustrate the practical potential of mathematical models in forensic data analysis.
Related Topics
Physical Sciences and Engineering
Chemistry
Analytical Chemistry
Authors
Anne-Laure Terrettaz-Zufferey, Frédéric Ratle, Olivier Ribaux, Pierre Esseiva, Mikhail Kanevski,