Modelling COVID-19 in Senegal and China with count autoregressive models

COVID-19 is a global health burden. We propose to model the dynamics of COVID-19 in Senegal and in China by count time series following generalized linear models. One of the main properties of these models is that they can detect potentials trends on the contagion dynamics within a given country. In particular, we fit the daily new infections in both countries by a Poisson autoregressive model and a negative binomial autoregressive model. In the case of Senegal, we include covariates in the models contrary to the Chinese case where the fitted models are without covariates. The short-term predictions of the daily new cases in both countries from both models are graphically illustrated. The results show that the predictions given by the negative binomial autoregressive model are more accurate than those given by the Poisson autoregressive model.