Evaluating data-driven methods for short-term forecasts of cumulative SARS-CoV2 cases

Background The WHO announced the epidemic of SARS-CoV2 as a public health emergency international concern on 30th January 2020. To date, it has spread to more than 200 countries and been declared global pandemic. For appropriate preparedness, containment, mitigation response, stakeholders policymakers require prior guidance propagation SARS-CoV2. Methodology This study aims provide such by forecasting cumulative COVID-19 cases up 4 weeks ahead for 187 countries, using four data-driven methodologies; autoregressive integrated moving average ( ARIMA ), exponential smoothing model ETS random walk forecasts RWF ) with without drift. these forecasts, we evaluate accuracy systematic errors Mean Absolute Percentage Error MAPE MAE respectively. Findings results show that methods outperform other two methods. Additionally, generate heat maps pictorial representation at risk having an increase in coming February 2021. Conclusion Due limited data availability during ongoing pandemic, less data-hungry short-term models, like , can help anticipating future outbreaks