A Justification for Applying the Principle of Minimum Relative Entropy to Information Integration Problems

SYNOPTIC ABSTRACTInformation integration activities typically occur in the early stages of decision making. The intent of these activities is to produce a clear picture of an uncertain situation by combining information from multiple sources. Bayesian inference may not be directly applicable at this stage of the decision process. This paper justifies the use of the principle of minimum relative entropy as an inference procedure for information integration. The justification provided establishes asymptotic equivalence between the minimum relative entropy result and the Bayes estimate with consistent information.