Probability-Possibility Transformation: - Application to Bayesian and Possibilistic Networks

Probability-possibility transformation is a purely mechanical transformation of probabilistic support to possibilistic support and vice versa. In this paper, we apply the most common transformations to graphical models, i.e., Bayesian into possibilistic networks. We show that existing transformations are not appropriate to transform Bayesian networks to possibilistic ones since they cannot preserve the information incorporated in joint distributions. Therefore, we propose new consitency properties, exclusively useful for graphical models transformations.