ACONVERGENCE OF MACHINE LEARNING AND STATISTICS TO PREDICT COVID-19 EVOLUTION

Abstract

The effect of the Covid pandemic is not restricted to sickness and death but also extends to socioeconomic concerns. The statistics-based assessment of covid data presented is the measure of the damages that happened to citizens of a country and the required actions taken towards those damages. This study aims to analyse the consequences of numerous considerations on the deaths due to the pandemic. The paper presents the statistics-based processing of COVID-19 data using logistic regression (LR) and decision tree (DT). Results are compared using the logistic regression algorithm (based on statistics) and the decision tree algorithm (based on machine learning). This study presented the predictive abilities of logistic regression and decision tree approaches and observed better results for the decision tree method. An accuracy of 94.10% for decision tree and 93.90% for logistic regression, respectively observed. It is also observed that highly populated countries are inclined to have more corona cases than those with low density. More females die than males, and a greater number of deaths are observed in cases of age greater than sixty-five years. The experimental data is gathered from the official website of the world health organization (WHO) between January 2020 and June 2020. The results presented are promising for the reported studies.