Identifying the risk factors for COVID-19 mortality in Sri Lanka using mutual information

Abstract

COVID-19 is a pandemic that has taken over the whole world. This virus can easily spread from person-to-person similar to the case with Severe Acute Respiratory Syndrome (SARS-CoV). COVID-19 affects people in different ways majority of the infected people will develop mild to moderate symptoms and recover without hospitalization. As of end of September, there were 516,465 cases and 12,906 deaths in Sri Lanka. The aim of this research is to identify the factors associated with the daily deaths due to COVID-19 in Sri Lanka. In the literature, there are many methods to find factors associated with COVID-19 deaths such as adjusted linear model, logistic regression analyses, and generalized additive model, etc. In this study, the Mutual Information (MI) based method was used to find the factors associated with daily death cases due to COVID-19. The MI was calculated using a k-nearest neighbors method.