Forecasting of COVID-19 Using Hybrid ARIMA-FB Prophet Algorithm

Abstract

Coronavirus is not a living organism. It has a genetic organism material covered with fat envelope. It affects throughout the country. The World Health Organization (WHO) officially announced that a new virus had been identified which then is called by 2019-nCoV on January 2020. The virus was recognized as part of the coronavirus group, which involves SARS and the other known colds. The first reported case was from Wuhan, China and has infected 7711 people and 170 reported deaths in China before coronavirus was declared as a global pandemic which produces a sickness authoritatively defined as COVID-19 that has diffused to a minimum 141 nations and regions, causing death over 5700 individuals around the world. The present paper explores the COVID-19 data frame seasonal wise and forecast future data for next 15 days in India using Auto Regressive Integrated Moving Average (ARIMA) and FB Prophet (Face Book Prophet) algorithm. The performance metrics used to find the model fitness are mean absolute error (MAE), mean percentage error (MPE) and root mean square error. Derived RMS of Prophet model as 39,407.98 for confirmed forecast COVID-19 in India date wise is simulated using Google Colaboratory to run and execute Python program.