Causality notion in the possibilistic framework has not been widely studied, despite its importance in context of poor or incomplete information. In this paper, we first propose an approach for handling interventions in quantitative possibilistic networks. The main advantage of this approach is its ability to unify treatments of both observations and interventions through the propagation process. We then propose a model based on quantitative possibilistic networks for ascribing causal relations between elements of the system by presenting some of their properties. Using such graphical structures allows to provide a more parci-monious inference process (comparing to the possibilistic model based on System P) that both accepts interventions and observations.
展开▼
