Conditional Approach to Possibility-Probability Fusion

Here, we are interested in the issue of multisource uncertain information fusion. We first consider the problem of fusing multiple possibility distributions. In this case, we particularly investigate the issue of normalization and suggest a generalized approach to the normalization of conflicting possibility distributions. We then look at the fusion of multiple probability distributions. We then suggest a framework to fuse probabilistic and possibilistic uncertainty based on the idea of conditioning the probability distribution with respect to the possibility distribution. We extend our approach to the case where our information about either of the distributions is imprecise, the so called second-order uncertainties.