A Statistical Viewpoint on the Theory of Evidence
IEEE Transactions on Pattern Analysis and Machine Intelligence
Evidential reasoning in a categorial perspective: conjunction and disjunction of belief functions
UAI'91 Proceedings of the Seventh conference on Uncertainty in Artificial Intelligence
| Hi-index | 0.00 |
Recently, Hummel and Landy proposed a variation on the Dempster/Shafer theory of evidence that tracks only the first and second order statistics of the opinions of sets of experts. This extension permits the tracking of statistics of probabilistic opinions, however, as opposed to tracking merely Boolean opinions (or possibilities within the "frame of discernment"). Both the Dempster/Shafer formulation and the Hummel/Landy formulation assume that bodies of experts that are combined to form new statistics have independent information. We give a model for parameterizing degree of dependence between bodies of information, and extend the Hummel/Landy formulation for combining evidence to account for sets of experts having dependent information sources.