Prediction of the Effectiveness of Government Measures towards Covid-19 Using Multiple Linear Regression Analysis

Abstract

Abstract
Several millions of people around the globe have been affected by the emergency of Coronavirus disease 2019 (COVID-19) pandemic. This menace has caused a geometrical mortality rate and stressed medical facilities in many counties. The lack of immediate treatment for the disease propelled the government of various countries to put in place some control measures to contain the rapid spread of the disease. Some of the measures taken include: staying at home orders, restricting movements, closing schools and workplaces, etc. This paper aims at determining the efficacy of these measures towards the containment of Covid-19. The Oxford Covid-19 Government Response Tracker (OxCGRT) dataset was used for this research. The data sets consist of daily entries of covid-19 cases in countries and various governments' active Covid-19 control measures/policies. Each policy indicator is given an ordinal score to denote its stringency. The data were analysed to gain insight into feature relationships and trends. Pearson's correlation coefficient (PCC) was used to select four features that contributed the most to the response variable. The training was carried out on 80% of the dataset using the python scikit-learn implementation of the Linear Regression algorithm, while testing was carried out on the remaining 20%. The model was trained on two features: The Containment Health Index which aggregates the other features, and the total number of Covid-19 cases. It achieved a coefficient of determination (r-squared) score of 0.09.