An Ensemble Learning Strategy for Panel Time Series Forecasting of Excess Mortality During the COVID-19 Pandemic

Abstract

Quantifying and analyzing excess mortality in crises such as the ongoing COVID-19 pandemic is crucial for policymakers. The traditional way it is measured does not account for differences in the level, long-term secular trends, and seasonal patterns in all-cause mortality across countries and regions. This paper develops and empirically investigates the forecasting performance of a novel flexible and dynamic ensemble learning strategy for seasonal time series forecasting of monthly respiratory diseases deaths data across a pool of 61 heterogeneous countries. The strategy is based on a Bayesian Model Ensemble (BME) of heterogeneous time series methods involving both the selection of the subset of best forecasters (model confidence set), the identification of the best holdout period for each contributed model, and the determination of optimal weights using the out-of-sample predictive accuracy. A model selection strategy is also developed to remove the outlier models and to combine the models with reasonable accuracy in the ensemble. The empirical results of this large set of experiments show that the accuracy of the BME approach improves noticeably by using a flexible and dynamic holdout period selection. Additionally, that the BME forecasts of respiratory disease deaths for each country are highly accurate and exhibit a high correlation (94%) with COVID-19 deaths in 2020.

Note:
Funding Information: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Keywords: Layered learning, Multiple learning processes, Time Series, Ensemble Bayesian Model Averaging (EBMA), SARS-CoV-2