Forecasting Time Series COVID-19 Statistical Data with Auto-Regressive Integrated Moving Average and Box-Jenkins' Models

Abstract

The current epidemic situation due to COVID-19 is a public health disaster worldwide. Forecasting play's, a crucial role in determining the pandemic's hypothetical situation and economic situation. It provides the base for authorities, public health officials, management teams, and other stakeholders to plan for future preventive actions in their companies, citizens, and governments. This paper proposes Auto-Regressive Integrated Moving Average mathematical modeling in integration with Box-Jenkins' model-building approach examining the variation in pandemic severity through the Loess smoothed curves to forecast the COVID-19 pandemic situation. The time-plot and forecasting results show Chinese resilience to pact with pandemic situation effectively whereas India was severely affected by the pandemic. The future forecast for India shows the worst situation by the end of 2021. Pakistan and Bangladesh are the least affected among the specified countries while decline in weekly death cases has been observed in Iran till the end of 2021. We observed the Case Fatality Ratio (CFR) of 2.08% globally.