Artificial Intelligence Forecasting of Covid-19 in China.

Background: Alternative to epidemiological models for transmission dynamics of Covid-19 in China, we propose the artificial intelligence (AI)-inspired methods for real-time forecasting of Covid-19 across China to estimate the size, lengths and ending time of Covid-19 across China. Method: We developed modified stacked auto-encoder for modeling the transmission dynamics of the epidemics. We applied it to real-time forecasting the lab confirmed cases of Covid-19 across China. The data were collected from January 19 to February 16, 2020 by WHO. We used the latent variables in the auto-encoder and clustering algorithms to group the provinces/cities for investigating the transmission structure. Results: We forecasted curves of cumulative confirmed cases of Covid-19 across China from January 20, 2020 to April 30, 2020. We used the multiple-step forecasting to estimate that the accuracies of one-step, 2-step, 3-step, 4-step and 5-step forecasting were -0.48%, 0.18%, 0.46%, 0.22% and 1.55%, respectively. We predicted that the time points of the provinces/cities entering the plateau of the forecasted transmission dynamic curves varied, ranging from February 10 to April 20, 2020. We grouped 34 provinces/cities into 9 clusters. Conclusions: The accuracy of the AI-based methods for forecasting the trajectory of Covid-19 was high. We predicted that the epidemics of Covid-19 will be over by the end of April. If the data are reliable and there are no second transmission, we can accurately forecast the transmission dynamics of the Covid-19 across the provinces/cities in China. The AI-inspired methods are a powerful tool for helping public health planning and policymaking.