Cluster analysis of countries based on healthcare indicators

This thesis examines the clustering of countries according to their level of healthcare based on systematically selected variables that view a country's health from many different perspectives. The aim is to uncover clusters of countries that are similar in their level of healthcare and to find how they are interlinked. The theoretical part is devoted to the description of the variables and the introduction of the transformations that are necessary for the subsequent clustering. At the same time, the clustering algorithms are introduced, namely hierarchical clustering, the method of k-means and Gaussian mixture models. In the practical part, these transformations and clusterings are performed on the data presented in the theoretical part of the thesis. As a result of the analysis, the presented clusters show that groups of different levels of healthcare seem to be present in the data. Moreover, these groups are similar between the algorithms and it is possible to find indicators that connect the countries in the clusters. Hierarchical clustering and Gaussian mixture models clustering seem to be the most appropriate solutions, since meaningful interpretations of the groups can be found for both solutions.