Mid-Term Forecasting of Fatalities Due to COVID-19 Pandemic: A Case Study in Nine Most Affected Countries

Rai, Sneha; De, Mala

The outbreak of COVID-19 pandemic has presented the entire world with an unrivalled challenge of public health leaving a remarkable impact on the social, economic, and financial lives of humanity. Though a major portion of the globe is under lockdown due to this deadly virus, the number of causalities is still growing rapidly. Therefore, it is very important to predict the number of infected and fatality cases for the future to overcome the consequences, save the lives of people, and plan accordingly. This paper proposes a data-driven analysis based on univariate Holt’s double exponential smoothening method with parameter optimization and polynomial curve fitting technique for one month ahead forecasting of the death cases in nine profoundly affected countries across the world namely India, USA, Italy, UK, China, France, Iran, Spain, and Germany. In contrast to the complex deep learning-based predictors, the proposed Holt’s model is simple yet efficient enough to give outstanding prediction performance for all the countries under this study and can be further used for the prediction of infections and causalities for the rest of the countries in future. The future estimation of the number of death cases will act as a beneficial tool for the successful allocation of the medical resources and as an early warning to the policymakers and health officials as well as to the residents of the country to boost their self-awareness.