Imprecise and Indefinite Probabilities

From AGIRI.org

Imprecise probabilities are due to Peter Walley, and involve representation of probabilities using intervals such as [.6,.8].

Indefinite probabilities are used in Probabilistic Logic Networks, and extend imprecise probabilities into probabilities of the form ([.6,.8],.9,4), with the meaning "the odds are .9 that after 4 more observations of the underlying phenomenon, the observed mean of the probability distribution will lie in the interval between .6 and .8". This is a highly compacted way of representing an envelope of probability distributions associated with a phenomenon.