Following the concept of intervals of measures, some general results are derived for the maximum imprecision related to imprecise densities. A model is proposed for Bayesian inference with imprecise prior densities, where members of a conjugate family can be chosen to enable updating in a simple way. Key Words: Bayesian inference, imprecise probabilities, intervals of measures, conjugate prior densities.
|Place of Publication||Eindhoven|
|Publisher||Technische Universiteit Eindhoven|
|Number of pages||13|
|Publication status||Published - 1992|