Fusion of Data and Expert Judgments with Imprecise Probabilities for Decision Making

Bayesian approaches to decision analysis are geared towards determining the optimal alternative by combining expert knowledge and data in small sample situations. However, Traditional Bayesian decision models assume that all probabilities are precise. In practical applications the probabilities of the states of nature are often under uncertainty due to the imprecision in the experts' judgments. In this paper, we present a decision support model that uses interval-valued probabilities to represent experts' uncertain beliefs. The model performs reasoning and decision making by integrating expert uncertain information and historical statistical data for Bayesian analysis. An example concerning risk assessment is given to demonstrate the applicability of the model in a real-world domain.