COVINet: A deep learning-based and interpretable prediction model for the county-wise trajectories of COVID-19 in the United States

ABSTRACT The cases of COVID-19 have been reported in the United States since January 2020. There were over 103 million confirmed cases and over one million deaths as of March 23, 2023. We propose a COVINet by combining the architecture of both Long Short-Term Memory and Gated Recurrent Unit and incorporating actionable covariates to offer high-accuracy prediction and explainable response. First, we train COVINet models for confirmed cases and total deaths with five input features, compare their Mean Absolute Errors (MAEs) and Mean Relative Errors (MREs) and benchmark COVINet against ten competing models from the United States CDC in the last four weeks before April 26, 2021. The results show that COVINet outperforms all competing models for MAEs and MREs when predicting total deaths. Then, we focus on the prediction for the most severe county in each of the top 10 hot-spot states using COVINet. The MREs are small for all predictions made in the last 7 or 30 days before March 23, 2023. Beyond predictive accuracy, COVINet offers high interpretability, enhancing the understanding of pandemic dynamics. This dual capability positions COVINet as a powerful tool for informing effective strategies in pandemic prevention and governmental decision-making.