The application of ARIMA model to analyze COVID-19 incidence pattern in several countries

The COVID-19 pandemic continues to spread and already shows a recurrence in many countries, despite several social distancing and vaccination measures implemented all around the world. Epidemiological data are available, and we use the Auto-Regressive Integrated Moving Average (ARIMA) model to analyze incidence pattern and to generate short-term forecasts of cumulative reported cases in Morocco, France, Italy, Spain and USA, using daily reported cumulative cases data from Worldometers, and we report 5-day and 10-day ahead forecasts of cumulative cases and check a posteriori the precision of this forecasting, by confronting it to the real data observed. In the discussion, we propose a link between the ARIMA, elevation and average temperature in several countries’ modelling approaches, for allowing the comparison between their explicative abilities.