Case-based belief formation under ambiguity (Q607257)

The paper contributes to the methods of decision making using randomized statistical experiments with repeated trials, and exploiting the experience and learning for improvement of the quality of accepted decisions. The main parts of the paper are devoted to extension and modification of previous knowledge of the problem (referred in the paper). The main result guarantees the uniqueness of so called similarity function over the resulting probability distributions characterizing the database of knowledge acquised during the learning process. The general theoretical methods and results are illustrated by several examples.