Machine Learning in Global Development: Applying k-Means Clustering to Identify Country Groupings by Economic and Health Performance

This study applies k-means clustering to group countries based on key economic and health indicators: Gross National Income per capita (GNIP), health expenditure, life expectancy, birth and death rates, and urbanization. The elbow method identified k = 3 as the optimal number of clusters, indicating a significant drop in within-cluster sum of squares (from 5000 to 2000). The results reveal three distinct development groupings. A small cluster of 13 high-performing countries stands out with strong economic (GNIP = 0.658) and health outcomes (Life Expectancy = 0.856), along with low birth (0.116) and death rates (0.113). This group also shows strong internal similarity (silhouette width = 0.58). The remaining countries fall into two broader clusters. The first includes 320 countries with moderate development, higher urbanization (UrbanP = 0.712), and relatively high health spending (healthE = 0.219), but lower GNIP (0.066). The second cluster of 277 countries faces greater challenges, marked by low life expectancy (0.414), high birth rates (0.670), and weak economic indicators (GNIP = 0.067). Both larger clusters show moderate cohesion (silhouette widths = 0.29 and 0.32). These findings highlight the stratified and multidimensional nature of global development, offering a data-driven framework to inform policy decisions and tailor interventions to the unique characteristics of each cluster.