Clustering of countries for COVID-19 cases based on disease prevalence, health systems and environmental indicators. (October 2021)

• Record Type: Journal Article
• Authors: Rizvi, Syeda Amna; Umair, Muhammad; Cheema, Muhammad Aamir
• Is Part Of: Chaos, solitons and fractals. Volume 151(2021)
• Publication Date: 2021-10
• Subjects: COVID-19 -- Clustering methods -- Unsupervised learning -- K-Means -- Second wave -- COVID-19 confirmed cases -- COVID-19 death cases -- Disease prevalence -- Pearson correlation; Chaotic behavior in systems -- Periodicals; Solitons -- Periodicals; Fractals -- Periodicals; Chaotic behavior in systems; Fractals; Solitons; Periodicals; 003.7
• DOI: 10.1016/j.chaos.2021.111240
• Languages: English
• ISSNs: 0960-0779

Abstract

Highlights: K-Means clustering of 79 countries has been performed for COVID-19 confirmed and death cases based on 18 feature variables. Correlation of all feature variables with COVID-19 confirmed cases and COVID-19 confirmed deaths have also been analysed. Asthma, diabetes mellitus, cardiovascular disease and nutritional deficiencies show positive correlation with COVID-19. It leads to useful insights related to a country's strategies that are impacting COVID-19 prevalence. Abstract: The coronavirus has a high basic reproduction number ( R 0 ) and has caused the global COVID-19 pandemic. Governments are implementing lockdowns that are leading to economic fallout in many countries. Policy makers can take better decisions if provided with the indicators connected with the disease spread. This study is aimed to cluster the countries using social, economic, health and environmental related metrics affecting the disease spread so as to implement the policies to control the widespread of disease. Thus, countries with similar factors can take proactive steps to fight against the pandemic. The data is acquired for 79 countries and 18 different feature variables (the factors that are associated with COVID-19 spread) are selected. Pearson Product Moment Correlation Analysis is performed between all the feature variables with cumulative death cases and cumulative confirmed cases individually to get an insight of relation of these factors with the spread of COVID-19. Unsupervised k-means algorithm is used and the feature set includes economic, environmental indicators and disease prevalence along with COVID-19 variables. The learning model is able to group the countries into 4 clusters on the basis of relation with all 18 feature variables. We also present an analysis of correlation between the selected feature variables, and COVID-19 confirmed cases and deaths. Prevalence of underlying diseases shows strong correlation with COVID-19 whereas environmental health indicators are weakly correlated with COVID-19.