Optimal Machine Learning Methods to Forecast COVID-19 Cases |Biomedgrid

The entire world is undergoing a hard-hitting scenario and trying to combat the COVID-19 by recent technological advancements, which involves machine learning chiefly. The forecasting demand has become a prerequisite as it helps the government officials and other organizations to make well-versed verdicts and impose pertinent measures to benefit the living conditions of the individuals around the globe. Consequently, the current paper focuses on four significant machine learning methods (Facebook Prophet, Auto Regression, Vector Auto Regression, and Holt-Winters), which help forecast the total confirmed and daily confirmed cases. Moreover, the paper reveals the ideal method for the future forecast based on the attained results and efficacy rate. The results of the study reveal the best methods for the considered countries based on the calculation of Error percentage. Out of the four Machine Learning models, AR and FB models stood out as the best methods when compared to the other two.