Representation of incomplete probabilistic information

Cédric Baudrit (1) , Didier Dubois (1, 2) , Hélène Fargier (1, 2)

1 IRIT-ADRIA - Argumentation, Décision, Raisonnement, Incertitude et Apprentissage
2 CNRS - Centre National de la Recherche Scientifique

This article deals with the representation of incomplete, vague knowledge and presents a nonexhaustive list of knowledge types which can be encountered when one makes risk evaluation calculations for man and environment. We will try to represent for each case the available information faithfully using different theories (possibility, probability, belief functions...). We will reconsider theoretical aspects such as the probability→possibility transformation and the probability families induced from possibility distributions.

Domaines: Intelligence artificielle [cs.AI]

Cédric Baudrit, Didier Dubois, Hélène Fargier. Representation of incomplete probabilistic information. Soft methodology and random information systems, 26, Springer, pp.149-156, 2004, Advances in Soft Computing book series (AINSC), 978-3-540-22264-4. ⟨10.1007/978-3-540-44465-7_17⟩. ⟨hal-02831349⟩