Modified Generalized Likelihood Uncertainty Estimation (GLUE) Methodology for Considering the Subjectivity of Likelihood Measure Selection

Abstract

The generalized likelihood uncertainty estimation (GLUE) methodology has been widely used in many areas as an effective and general strategy for model calibration and uncertainty estimation associated with complex models. The application of GLUE requires a formal definition of a likelihood measure. However, it has been recognized that the choice of a likelihood measure is inherently subjective. This, in turn, introduces a new kind of uncertainty—the uncertainty owing to the lack of knowledge in choosing the true likelihood measure in the GLUE methodology. This study proposes a practical framework to address this uncertainty by using multiple likelihood measures, analogous to considering multiple expert opinions. The final uncertainty probability estimates are then obtained by combining the estimates from individual likelihood measures based on probability theory.