Real-time short-term forecast of COVID-19 epidemic in Cuba using model averaging

Abstract

In the absence of reliable information about transmission mechanisms of an emerging infection, simple phenomenological models can provide an early assessment of the potential scope of outbreaks in near real- time. Early prediction of the final size of any epidemic and in particular for ongoing COVID-19 epidemic can be useful for health authorities in order to plan the response to the outbreak. A variety of nonlinear models have been developed to model reported cumulative cases in infectious disease outbreak (e.g., Richards, logistic, Gompertz models). All these models could fit epidemic data well in order to obtain real-time short-term forecasts. Typically, one follows the so called post selection estimation procedure, i.e., selects the best fitting model out of the set of candidate models and ignores the model uncertainty in both estimation and inference since these procedures are based on a single model. In this paper, we conduct a real-time prediction for the final size, turning point of the outbreak, and also generate 10-day ahead forecasts of cumulative case using several nonlinear models in which these parameters are estimated via model averaging. The proposed method is applied to COVID-19 epidemic data in 2020 Cuba outbreak.