Comparison of K-means, Gaussian Mixture, and Hierarchical Clustering Models on Countries' Economic Freedom Index

Economic freedom scores can be used as a measure to assess prosperity and opportunity. This article provides a clustering analysis of the countries based on the Economic Freedom Index data published by the Heritage Foundation. K-means, gaussianmixture model, and hierarchical clustering are used. The research objective is to find an appropriate number of clusters as a hyperparameter for the methods. Then, for identifying a good clustering method for clustering the countries. We use scaling and PCA for dimensionality reduction for preprocessing the data. The resulting best clustering method among those used is the gaussian mixture model, with three resulting clusters. This grouping seems inline when compared to the overall score quartiles of the data. Thetwo other two methods used, k-means and hierarchical cluster analysis, also generally group the data properly, with four and two clusters respectively. However, there are noticeable anomalies when compared to the quartiles of overall score.