Modeling the Omicron Dynamics and Development in China with a Deep Learning Enhanced Compartmental Model

Background: Compartmental models dominate epidemic modeling. Estimations of transmission parameters between compartments are typically done through stochastic parameterization processes that depend upon detailed statistics on transmission characteristics, which are economically and resource-wide expensive to collect.ObjectivesWe apply deep learning techniques as a lower data dependency alternative to estimate transmission parameters of a customized compartmental model, for the purpose of simulating the dynamics of the Omicron phase of the COVID-19 epidemics and projecting its further development in China and subregions within the country.Methods: We construct a compartmental model, and develop a multivariate, multistep deep learning methodology to estimate the model’s transmission parameters. We then feed the estimated transmission parameters to the compartmental model to predict the development of the COVID-19 epidemics in China and subregions within the country for 28 days.Results: In China (excluding Hong Kong and Taiwan), the daily Omicron infection increase is between 60 and 260 in the 28-day forecast period between June 4 and July 1, 2022. On July 1, 2022, there would be 768,622 cumulative confirmed cases and 591 cumulative deceased cases. The CFR would stabilize at 0.077%±0.00025%. Assuming a 25% infection rate, the total deaths with Omicron would be up to 280,000 without non-pharmaceutical intervention (NPI).Conclusions: Current compartmental models require stochastic parameterization to estimate the transmission parameters. These models’ effectiveness depends upon detailed statistics on transmission characteristics. As an alternative, deep learning techniques are effective in estimating these stochastic parameters with greatly reduced dependency on data particularity.