A Comparative Analysis of Machine Learning Techniques for Exploring Country Clustering Based on Life Expectancy

This study employs data mining techniques and clustering algorithms to analyze life expectancy factors and discover influential elements. Using World Health Organization data, including life expectancy figures for various countries and years, the study examines features such as mental and physical health conditions, disease rates, and accidents. Three clustering models—DBSCAN, k-means, and fuzzy c-means—are developed and evaluated using metrics such as Silhouette score, Davis Bouldin Index (DBI), and Calinski-Harabasz index. The results favor the k-means model, with a Silhouette coefficient of 0.7812, a Calinski-Harabasz score of 541.2260, and a Davies-Bouldin score of 0.3297. Compared to previous studies, this k-means model excels, highlighting its effectiveness in analyzing life expectancy factors. The findings contribute to global efforts in improving life expectancy and offer insights for policymakers, UNDP, and other stakeholders. Identifying key factors can guide targeted interventions, resource allocation, and policy development to enhance healthcare infrastructure and address poverty in underprivileged countries. This research emphasizes data-driven decision-making and inspires collaborative efforts for a healthier and more equitable world.