Generalizing Markov Chains to Imprecise Previsions

摘要

The requirement that precise probabilities be available is often not realistic in many applications. Some other consistent probabilistic approaches are needed to allow for imprecision in statistical reasoning. In this paper the theory of coherent imprecise previsions is employed to generalise discrete Markov chains to imprecise probabilities. A general procedure of imprecise probability elicitation is analysed as well. An example is provided.