Bayesian Networks for Probabilistic Inference and Decision Analysis in Forensic Science, 2nd Edition

7.9 Kinship analyses

7.9.1 A disputed paternity

The application of graphical probabilistic models, notably Bayesian networks, to inference problems involving the results of DNA profiling analyses represents a lively area of research. A particular application is kinship analyses as discussed by Dawid et al. (2002). These authors discuss how to derive appropriate graphical structures for Bayesian networks from initial pedigree representations of forensic identification problems. This section focusses on such an example in terms of a classic case of disputed paternity.

A certain male, denoted as the putative father , is alleged to be the father of a certain child c. Results are available of measurements taken on genetic markers for the mother m, the putative father and the child c. What is thought is the likelihood ratio for the proposition that the putative father is the true father , noted , given knowledge about the child's genotype and the mother's genotype , when the alternative ...