Evaluating Countries According to Global Risk Index: A Clustering Analysis Application

Abstract

Purpose – In this study; Global Risk Index for 191 countries were taken into consideration and a data mining study was conducted on cluster analysis by evaluating the negativities faced by societies. The aim of the study is to ensure that countries are grouped with cluster analysis, to rank the clusters in terms of risk and to determine which cluster countries are in in terms of risk. Design/Mothodology/Approach – In this study, clustering was performed with the Expectation Maximization Algorithm, and then it was examined whether there was a significant difference on the basis of variables with the Kolmogorov-Smirnov Normality Test, the Levene Statistic Test of Homogeneity of Variances and the Kruskal-Wallis Significance Test. The 18 variables used in the study were divided into 10 clusters for 2019, taking into account the literature review and expert opinions, and the clusters were evaluated in terms of risk. Findings – The Expectation Maximization Algorithm reveals significant results in the clustering analysis and when the clustering results are analyzed, risky and risk-free countries are seen in terms of global risk indicators. Discussion – In this context, it is thought that countries can be helped to take the necessary precautions regarding risk by making their own evaluations. Covid-19, which is experienced today and continues to affect the world, is also a global risk. It has been seen how unprepared the countries all over the world are for such risks.