Use of Bayesian Networks for the Forensic Evaluation of Gunshot Residue

Gunshot residue is a common type of trace considered in forensic science. It contains information about a past event, typically a firearm incident. As with any forensic trace, meaning must be assigned to it. However, from the moment it leaves the firearm to its recovery from a surface of interest, gunshot residue undergoes dynamic activities that pose challenges to its interpretation. The evaluative reporting framework is commonly reported as a valuable approach to assist with this part of the forensic science process. This review presents some major research on the use of Bayesian networks and the likelihood ratio approach to gunshot residue evaluation. The benefits and challenges of statistical modeling are discussed. A mock scenario is used to illustrate how a Bayesian network might be built. Sensitivity analysis and a multivariate approach are identified as valuable future research directions in this field.