Sets of probability measures which form neighbourhoods of additive probability measures are studied. An application of the Jeffrey's rule of conditioning to forming neighbourhoods of probability measures is proposed. Neighbourhoods that are closed for generalized conditioning according to this rule are characterized. They are shown to be exactly convex and bi-elastic neighbourhoods.
Keywords. interval probability, robust statistics, non-additive probabilities, neighbourhood models
The paper is availabe in the following formats:
Kardeljeva ploscad 5