A COMPARATIVE ANALYSIS OF DBSCAN AND GAUSSIAN MIXTURE MODEL FOR CLUSTERING INDONESIAN PROVINCES BASED ON SOCIOECONOMIC WELFARE INDICATORS

Public welfare refers to a condition in which people experience happiness, comfort, prosperity, and can adequately fulfill their basic needs. Indonesia consists of several provinces, each with varying levels of welfare. One crucial aspect in promoting equitable development is ensuring that all regions in Indonesia achieve similar welfare standards. This study aims to classify Indonesian provinces based on socioeconomic welfare indicators, with the results serving as a basis for policy-making that considers regional potential and challenges. The data used in this study are secondary data obtained from the official website of BPS-Statistics Indonesia on provincial welfare indicators from 2020 to 2023. The research methodology includes data collection, descriptive statistical analysis, determining the optimal number of clusters, and comparing the clustering performance of Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and the Gaussian Mixture Model (GMM) using Silhouette Index, Davies-Bouldin Index, and Calinski-Harabasz Index as evaluation metrics. The DBSCAN-based clustering resulted in two clusters: high-welfare and low-welfare regions. Meanwhile, GMM clustering produced five clusters: moderate, fairly low, low, high, and fairly high welfare regions. Based on cluster validity measures, GMM outperformed DBSCAN, achieving a Silhouette score of 0.28, a Davies-Bouldin Index of 1.12, and a Calinski-Harabasz Index of 10.9.