ABSTRACT
Expert judgment is a critical component of forecasting methods and can be implemented by a variety of forecasting tools. Expert judgment has long been employed in computational modeling to develop and structure Bayesian networks (BNs). As the result of advanced computational algorithms that enabled BNs to learn from data, modelers increasingly find expert judgment useful, and even crucial in some cases, in dealing with data limitation problems and calibrating model parameters (Walsh et al. 2010; Henrion et al. 1991). Due to their technical simplicity and methodological transparency, we focused on four forecasting methods regarding their linkability with Bayes net models: conjoint analysis, probability elicitation, judgmental bootstrapping, and prediction markets. We recognize that different decision-making environments and conditions undergird each of these methods, and their application is largely influenced by the characteristics of the research questions at hand. We do not propose to incorporate all existing judgmental forecasting methods into the integrative framework, or identify the most superior forecasting method, which would be beyond the scope of this project. Our goal is to shed some exploratory light on integrative modeling with the hope of ushering in additional research interests in modeling complex decision-making and risk-assessment problems.
