Handling manipulated evidence


Bayesian Networks have been advocated as useful tools to describethe relations of dependence/independence among random variables andrelevant hypotheses in a crime case. Moreover, they have been appliedto help the investigator structure the problem and evaluate the impactof the observed evidence, typically with respect to the hypothesisof guilt of a suspect. In this paper we describe a model to handlethe possibility that one or more pieces of evidence have been manipulatedin order to mislead the investigations. This method is based on causalinference models, although it is developed in a different, specificframework.

Forensic Science International
Gianluca Baio
Gianluca Baio
Professor of Statistics and Health Economics