Uncertainty Quantification and Global Sensitivity Analysis in Pandemic Modelling
by Emanuele Borgonovo, Xuefei Lu, Alessia Melegaro
Quantifying uncertainty in the predictions of quantitative models for epidemics forecasting is an important task, made even more urgent by the current COVID-19 pandemic. On the one hand, uncertainty in the forecasts needs to be properly represented to decision-makers. On the other hand, being able to explain such uncertainty identifying the key-drivers of uncertainty is an important task for the modeler and the analysts. On the one hand for information collection and, on the other hand, to compare the output of alternative models. This research project aims at addressing the challenges for making use of the advances in global sensitivity analysis methods in the quantification and decomposition of uncertainty in epidemic modelling considering the alternative nature of the models, their computational cost, the type of response and the goals of the models themselves.