Forcasting the Number of Infections and Deaths due to Alpha and Delta Variants of COVID-19 using ARIMA and Prophet Models

COVID-19 is an infectious disease caused by the most recently discovered Coronavirus. Despite four years of study on the characteristics of this virus since its emersion in 2019, lots of questions have been left unsolved. In this study, we extend examining the predictive power of statistical models in capturing dynamics of COVID-19 in Alpha and Delta variants in the three geographical regions of country of Iran, continent of Asia, and the world. We use the models of ARIMA and Prophet to predict the number of infections and deaths due to the aforementioned variants. The results show that ARIMA is superior to Prophet in the majority of predictions. This model is further superior to the newly LSTM models proposed for predictions in the same situation. We also provide a simulation of the spread of COVID-19 pandemic in Alpha and Delta variants in a test environment, to further illustrate the differences in the nature of the two variants.