Quick and Accurate Forecasting of COVID-19 Cases using Prophet Model

Abstract

This World requires the support of new technologies like Artificial Intelligence (AI), to face and look ahead the new problems, especially against serious threats like COVID-19. The objective of this proposed work is to utilize the artificial intelligence and its prominent discipline, Machine Learning (ML), as a decisive technology to fight with corona virus disease. In this work, Prophet Model, an open source tool developed by Face book is used to forecast the positive tested cases on a given particular day over Indian population. The proposed model is trained with existing data to find patterns to make forecast about speeding of virus in near future. The model validated with an accuracy of 96.6 % and forecasted the total number of cases for 30 days from June 28^th 2020 to July 28^th 2020, where the forecasted total number of cases in India reaches 857668 on the 30^th day, i.e. by July 28^th 2020.