From imprecise to granular probabilities (Q2386233)

The author discusses G. de Cooman's work on imprecise probabilities. First he points out that imprecise probabilities in de Cooman's sense are a special case of granular probabilities in the author's sense. Second, the class of imprecise probabilities in de Cooman's sense is not closed under imprecisely defined operations, i.e. if initial probabilities are imprecise, the induced probabilities are, in general, granular. And third, computations of granular probabilities requires concepts and techniques drawn from the theory of generalized constraints.