Comparison and Evaluation of Euclidean Distance and Divergence in Adaptive K-Means Algorithm for Clustering Human Development Index of Indonesia Province

This research explores the application of the Adaptive K-Means clustering algorithm on Human Development Index (HDI) data across 34 provinces in Indonesia, comparing the performance of Euclidean and Divergence distance metrics. The HDI indicators used include life expectancy, years of schooling, and per capita expenditure. Data processing was conducted both manually on sample data and automatically using Python for the complete dataset. Results demonstrate that the choice of distance metric significantly impacts clustering effectiveness. Divergence outperformed Euclidean based on silhouette score evaluations, offering more representative cluster separation. Scatter plot visualizations tracked the iterative clustering process. The study contributes to optimizing clustering techniques for socio-economic indicators such as HDI.