Generalized Bayesian inference in a fuzzy context: from theory to a Virtual Reality application

A generalized Bayesian inference framework in order to embed fuzzy sets and partial probabilistic information is provided. The general framework of reference is that of coherent conditional probabilities, which allows giving a rigorous interpretation of membership function as a conditional probability, regarded as a function of the conditioning event. The inferential problem needs to be studied in situations where the prior can be partial, moreover membership and prior can be given on different classes of events. This inferential model is applied to the virtual representation of a female avatar.