Quantifying the Value of Forensic Handwriting Evidence using Open-Source Feature Extraction
Gabrielle Collins · 2022
Handwriting documents, as forensic evidence, can be analyzed statistically to assist forensic document examiners in establishing an association between a document and a specific writer. This type of identification problem has been investigated previously by scientists at the Center for Statistics and Applications of Forensic Evidence (CSAFE), and the goal of this analysis was to extend their work by using an open-set analysis, where complete knowledge of some background population is not needed. This was completed and evaluated using both a Bayes Factor and likelihood ratio. Ultimately, both analyses supported the correct writer identification, given two separate populations of writers. The sampling models established and analysis used produced accurate results and should be considered in the evaluation of handwriting evidence in the future.