14:45 - 15:05 Methods for computing score-based likelihood ratios (SLRs) with dependent data

Abstract

Score-based likelihood ratios (SLRs) are a statistical method used to quantify the weight of evidence and have been used in many forensic disciplines. Using the context of footwear evidence to make the problem more concrete, similarity/dissimilarity scores are computed from pairs of shoeprints. Therefore, there is dependence between scores that share a shoeprint or shoe in common. Given up to 5 similarity/dissimilarity scores, we include four different methods of computing the SLR that address the dependence. We also provide a visualization of the resulting SLR using an ROC curve. More information can be found on the corresponding Github repository.