All Models Are Useful: Bayesian Ensembling for Robust High Resolution COVID-19 Forecasting

Speaker: Aniruddha Adiga, Research Scientist within the Network Systems Science and Advanced Computing division at the University of Virginia's Biocomplexity Institute

Abstract: Timely, high-resolution forecasts of infectious disease incidence are useful for policymakers in deciding intervention measures and estimating healthcare resource burden. Although multiple methods have been explored for this task, their performance has varied across space and time due to noisy data and the inherently dynamic nature of the pandemic. In this talk, we discuss our efforts towards forecasting COVID-19 confirmed cases that have continued to serve local, state, and multinational health agencies for over a year. We present a forecasting pipeline that incorporates probabilistic forecasts from multiple statistical, machine learning, and mechanistic methods through a Bayesian ensembling scheme and is able to produce forecasts at different resolutions: county, state, and national level. In terms of forecast evaluation, while showing that the Bayesian ensemble is at least as good as the individual methods, we also show that each individual method contributes significantly for different spatial regions and time points. We compare our model's performance with other similar models being integrated into CDC-initiated COVID-19 Forecast Hub. In addition, we have developed a public-access interactive dashboard for visualizing and evaluating our model forecasts and will discuss some of its functionalities.

Join this Zoom webinar here.