<![CDATA[Pneumonia is one of the leading causes of morbidity and mortality among children under five in Indonesia, including in West Java Province, which records the highest number of childhood pneumonia cases nationally. The spatial variation in pneumonia incidence indicates substantial heterogeneity across regions that cannot be adequately captured by global regression models. This study aims to analyze the factors influencing the number of pneumonia cases among children under five in West Java in 2024 using the Mixed Geographically Weighted Regression (MGWR) approach. Data were obtained from the West Java Provincial Health Office and Statistics Indonesia (BPS), comprising the number of pneumonia cases and predictor variables including exclusive breastfeeding coverage, poverty rate, percentage of low birth weight (LBW) infants, PHBS (clean and healthy living behavior), complete basic immunization coverage, and adequate housing. The global linear regression results show that not all variables have significant effects, and the model fails to adequately capture data variation, with indications of heteroskedasticity—thereby necessitating a spatial modeling approach. While the GWR model identifies local variation, model comparison demonstrates that the MGWR model performs best, yielding a lower AIC value (474.46) and a higher coefficient of determination (69.83%) than both GWR and the global model. MGWR identifies LBW and immunization coverage as global variables with consistent effects across all areas, whereas exclusive breastfeeding, PHBS, poverty, and adequate housing exhibit spatially varying influences. Cluster analysis using MGWR identifies five regional groups, each reflecting distinct determinants of childhood pneumonia in West Java. These findings highlight that public health interventions aimed at reducing childhood pneumonia must incorporate local spatial conditions, and MGWR provides a more appropriate analytical approach than global regression models.]]>
