Two-stage approach for the inference of the source of high-dimension and complex chemical data in forensic science

While scholars advocate the use of a Bayes factor to quantify the weight of forensic evidence, it is often impossible to assign the necessary probability measures for high-dimension and complex data, and so performing likelihood-based inference is impossible. We address this problem by leveraging the properties of kernel functions to propose an inference framework based on a two-stage approach to offer a method that allows to statistically support the inference of the identity of source of trace and control objects. Our method is generic and can be tailored to any type of data encountered in forensic science or pattern recognition.