Date of Publication :20th August 2020

Abstract: This study analyzes COVID-19 data and predicts the recovered cases in the US using machine-learning models. We employ three most common base learners, namely, linear regression, MLP and SVM, for analyzing and predicting COVID-19 recovered cases. Utilizing these linear regression and SVM base learners, we predict the recovered cases that are very satisfying and close to the exact recovered case. However, the MLP model predicted values are lower than that of actual values because of the low number of recovered cases data available. The predicted values are compared for their accuracy in predicting future recovery cases in the US. The linear regression model is found to be most accurate in comparison to the other models in this study. The key findings of this study will help in understanding the trend in coronavirus spread in the US and its recovery rate. It can help by providing a piece of valuable information to healthcare authorities and workers to design appropriate strategies for reducing the death toll and understanding the recovery rate of coronavirus patients in the US.

Reference :