Deep Forecasting of COVID-19: Canadian Case Study

COVID-19 is an infectious disease, which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). In this research, we firstly present an overview of the main forecasting models to predict the new cases of COVID-19. In this context, we focus on univariate time series models to analyze the dynamic change of this pandemic through time. We also introduce multivariate time series forecasting using weather and daily tests data, to study the impact of exogenous features on the progression of COVID-19. Finally, we present an ensemble learning model based on LSTM and GRU and evaluate the results using the MAE, RMSE and MAPE. The results show that the ensemble approach performs well in comparison to other models. In addition, this research provides an outcome regarding the dynamic correlation between temperature, humidity and daily test data and its impact on the new reported cases of contamination. © 2021, Springer Nature Switzerland AG.