DeepGLEAM: A hybrid mechanistic and deep learning model for COVID-19 forecasting

TL;DR

DeepGLEAM is a hybrid COVID-19 forecasting model that combines mechanistic simulations with deep learning to improve accuracy and uncertainty quantification.

Contribution

It introduces a novel hybrid approach that leverages mechanistic models and deep learning to enhance COVID-19 forecasting performance.

Findings

Improved mortality forecasting accuracy.

Effective uncertainty quantification with confidence intervals.

Demonstrated success on real-world COVID-19 data.

Abstract

We introduce DeepGLEAM, a hybrid model for COVID-19 forecasting. DeepGLEAM combines a mechanistic stochastic simulation model GLEAM with deep learning. It uses deep learning to learn the correction terms from GLEAM, which leads to improved performance. We further integrate various uncertainty quantification methods to generate confidence intervals. We demonstrate DeepGLEAM on real-world COVID-19 mortality forecasting tasks.