Speaker: Danica Ommen, Iowa State University, "Statistical Analysis of Handwriting for Forensic Identification of Source"

Abstract: Handwriting analysis is conducted by forensic document examiners who can visually recognize characteristics of writing to assess who wrote a questioned document. Recently, there have been incentives to investigate how to make this process less subjective and to increase the scientific and statistical rigor of the process. To this end, we use an automatic algorithm within the open-source ‘handwriter’ package in R to segment a handwritten sample into small graphical units of writing. These graphs are sorted into exemplar groups or clusters. We assume that the frequency with which a writer produces graphs to each cluster is characteristic of their overall writing profile. We address the question of who wrote a particular document using a number of different statistical techniques, including Bayesian classification and statistical learning methods, with the goal of quantifying the weight of the handwriting evidence to determine its source.