SOM Clustering of OECD Countries for COVID-19 Indicators and Related Socio-economic Indicators

Abstract

The coronavirus disease is one of the most severe public health problems globally. Governments need policies to better cope with the disease, so policymakers analyze the country's indicators related to the pandemic to make proper decisions. The study aims to cluster OECD (Organisation for Economic Co-operation and Development) countries using COVID-19, health, socioeconomic, and environmental indicators. A self-organizing map (SOM) clustering method, an unsupervised artificial neural network (ANN) method and a hierarchical clustering method are used. The data comprises 38 OECD countries, and 16 different variables are selected. As a result, the countries are grouped into 3 clusters. Cluster 1 contains 33 countries, the USA is Cluster 2, and Cluster 3 has 4 countries, including Turkey. COVID-19 mortality is highly related to mortality from chronic respiratory diseases. In addition, environmental indicators show differences in clusters.