A new method for aggregating experts' probability judgments

This paper proposes a new method to aggregate experts' probability judgments based on judgment evidence. The method uses a single causal Bayesian Network to express the judgment evidence of a single expert, and the final judgments are demonstrated through the integration of multiple causal Bayesian Networks into a single one, with the technology of reversal of arcs and the introduction of latent variables. This paper demonstrates the application of this method and raises some problems that occurs during its performance, by using a case of risk analysis for a large scale ground test of Air-to-Surface Missile.