Geographically weighted regression with Burr XII distribution for spatial analysis of diarrhea incidence in Surakarta City

Abstract

Diarrhea is one of the leading causes of mortality in Indonesia. The persistence of this disease is closely related to spatial factors, such as environmental conditions and existing infrastructure. In Surakarta City, diarrhea was recorded as the most prevalent disease in 2024, with a 58% increase in cases compared to the previous year. The aim of this study is to apply the Geographically Weighted Regression (GWR) approach with the Burr XII distribution to model the spatial distribution of diarrhea cases in Surakarta. The study encompasses 54 urban villages and utilizes seven predictor variables, such as population density, elevation, distance to the nearest hospital, slope, proximity to rivers, distance to waste disposal sites, and rainfall. The Burr XII distribution was applied to handle the skewness frequently observed in health-related datasets such as diarrhea incidence, while GWR was used to generate location-specific parameter estimates. The developed model revealed clear spatial heterogeneity in how diarrhea cases are influenced by the predictor variables. Several factors, including population density, hospital distance, elevation, and river proximity, showed varying positive and negative effects depending on the geographical area. The results underline the necessity for targeted and well-coordinated public health measures to mitigate diarrhea risk, especially in communities with inadequate healthcare services and poor waste management systems.