An extension of the conditional probability concept to Random-Fuzzy Variables representing measurement results

In measurement applications, the concept of conditional probability is used to evaluate the probability density function associated to a measurement result (and its associated uncertainty) when <I>a priori</I> knowledge about the measurand is available. However, in the recent years, different approaches than the theory of probability have been proposed to model measurement uncertainty, based on the possibility theory. Within this theory, a measurement result is expressed as a random-fuzzy variable in terms of possibility distributions. This paper aims at proposing a method for extending the conditioning concept to measurement results expressed in terms of random-fuzzy variables.