A Continuous Bayesian Model for the Stimulation COVID-19 Epidemic Dynamics

It is of great theoretical and application value to accurately forecast the spreading dynamics of COVID-19 epidemic. We first proposed and established a Bayesian model to predict the epidemic spreading behavior. In this model, the infection probability matrix is estimated according to the individual contact frequency in certain population group. This infection probability matrix is highly correlated with population geographic distribution, population age structure and so on. This model can effectively avoid the prediction malfunction by using the traditional ordinary differential equation methods such as SIR (susceptible, infectious and recovered) model and so on. Meanwhile, it would forecast the epidemic distribution and predict the epidemic hot spots geographically at different time. According to the results revealed by Bayesian model, the effect of population geographical distribution should be considered in the prediction of epidemic situation, and there is no simple derivation relationship between the threshold of group immunity and the virus reproduction number R0. If we further consider the virus mutation effect and the antibody attenuation effect, with a large global population spatial distribution, it will be difficult for us to eliminate Covid-19 in a short time even with vaccination endeavor. Covid-19 may exist in human society for a long time, and the epidemic caused by re-infection is characterized by a wild-geometric && low-probability distribution with no epidemic hotspots.