OOBN for Forensic Identification through a searching in a database of DNA profiles’

In this pap er we evaluate forensic identication hyp otheses conditionally to the characteristics observed b oth on a crime sample and on individ- uals contained in a database. First we solve the problem via a computational ecient Bayesian Network obtained by transforming some rec- ognized conditional sp ecic indep endencies into conditional indep endencies. Then we prop ose an Ob ject Oriented Bayesian Network represen- tation, rst considering a generic characteristic, then inheritable DNA traits. In this resp ect we show how to use the Ob ject Oriented Bayesian Network to evaluate hyp otheses concerning the p ossibility that some unobserved individuals, ge- netically related to the individuals proled in the database, are the donors of the crime sample.