Fachri Faisal, Henny Pramoedyo, Suci Astutik, Achmad Efendi
This research investigates the utilization of Bayesian Geographically Weighted Regression (BGWR) combined with Kriging to improve spatial predictions of the Human Development Index (HDI) throughout Sumatra Island. The results underscore considerable geographical variations in human development indicators, such as life expectancy, education, and economic measures. While the standard Geographically Weighted Regression (GWR) model produces reliable results, incorporating Bayesian frameworks—particularly with Jeffreys' uninformative prior—delivers superior predictive performance and effectively addresses spatial heterogeneity. The Bayesian Jeffreys model achieves the highest accuracy across multiple metrics, including the lowest mean absolute bias (MAB) and root mean squared error (RMSE), explaining 99.99% of HDI variance. Moreover, the Jeffreys model specifically reduces spatial autocorrelation in residuals, lowering the demand for further methods, including Kriging. By proving the superiority of Jeffreys' previous over conjugate priors in handling spatial heterogeneity and enhancing prediction accuracy, this work adds to the body of knowledge. It provides a strong basis for precision-driven spatial analyses in regional development planning and resource allocation. The results also highlight the urgent need for focused policy interventions to solve ongoing disparities in underdeveloped areas using metropolitan centers as scalable development models. © 2025 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license
Department of Mathematics, Faculty of Mathematics and Natural Sciences, Brawijaya University, Malang, 65145, Indonesia; Department of Mathematics, Faculty of Mathematics and Natural Sciences, Bengkulu University, Bengkulu, 38371, Indonesia; Department of Statistics, Faculty of Mathematics and Natural Sciences, Brawijaya University, Malang, 65145, Indonesia
