COVID-19 Pandemic Prediction using Time Series Forecasting Models

TL;DR

This paper compares ARIMA and Prophet time series models to forecast COVID-19 cases globally and in top affected countries, finding ARIMA more effective for predicting the pandemic's spread.

Contribution

It introduces a comparative analysis of ARIMA and Prophet models for COVID-19 case prediction using global and country-specific data from early 2020.

Findings

ARIMA outperforms Prophet in accuracy metrics

Forecasting models can inform government policy planning

Analysis enhances understanding of COVID-19 spread trends

Abstract

Millions of people have been infected and lakhs of people have lost their lives due to the worldwide ongoing novel Coronavirus (COVID-19) pandemic. It is of utmost importance to identify the future infected cases and the virus spread rate for advance preparation in the healthcare services to avoid deaths. Accurately forecasting the spread of COVID-19 is an analytical and challenging real-world problem to the research community. Therefore, we use day level information of COVID-19 spread for cumulative cases from whole world and 10 mostly affected countries; US, Spain, Italy, France, Germany, Russia, Iran, United Kingdom, Turkey, and India. We utilize the temporal data of coronavirus spread from January 22, 2020 to May 20, 2020. We model the evolution of the COVID-19 outbreak, and perform prediction using ARIMA and Prophet time series forecasting models. Effectiveness of the models are evaluated based on the mean absolute error, root mean square error, root relative squared error, and mean absolute percentage error. Our analysis can help in understanding the trends of the disease outbreak, and provide epidemiological stage information of adopted countries. Our investigations show that ARIMA model is more effective for forecasting COVID-19 prevalence. The forecasting results have potential to assist governments to plan policies to contain the spread of the virus.