Supervised Machine Learning Approach for the Prediction of COVID-19 Cases

ABSTRACT

COVID-19 was first discovered in the city of Wuhan. From then onward the virus has spread rapidly infecting thousands of people. The virus is still spreading and attempts are being made to predict and control the growth and spread of this virus. The trend of spread of this virus is highly unpredictable and normal statistical methods of predictions have not provided promising results, thus another approach of predicting the growth of this virus is required. This approach must be able to predict the nonlinear growth of the virus. Thus, an attempt is made to predict the growth of this virus and to show that the normal statistical methods are not able to predict the growth of the virus with high accuracy. The linear predicting algorithms used are Linear Regression, Support Vector Machine, Polynomial Regression and Auto Regressive Integrated Moving Average. The nonlinear predicting algorithm used is Prophet Algorithm for the prediction of exponential growth of spread of the virus. A comprehensive study is done to show how the spread of the virus takes place in different countries. A comparative study is also done to show the differences in performance parameters based on Absolute Mean Error, Mean Squared Error and R-squared (R2) score among different types of predictors. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.