Graph mining for forensic databases

Abstract

This paper provides an important proof of concept study for collecting and interrogating ballistic data used in informing forensic data investigations in territories like these. The work is grounded in the well-established forensic model frameworks for digital investigations. The experimental analysis provided in this study uses R studio to simulate and interpret the results collected against a NO-SQL graph database called Neo4J. We then apply graph mining techniques to determine levels of precision and accuracy against the compiled ballistic data sets (i.e. firearm data, and associated gang like activities related to such firearms). The findings of this study provide motivational case arguments to support law enforcement and the judicial process with regards to intelligence-led policing and prosecution by the courts.