<![CDATA[Malnutrition among infants and toddlers accounts for 45% of global child deaths, making accurate risk modeling essential. In Central Java, data on malnutrition cases show overdispersion (ϕ = 162.58), rendering the Poisson Regression (PR) model invalid. This study applied Generalized Poisson Regression (GPR) with Fisher-Scoring optimization to five predictor variables: complete basic immunization, active community health posts (posyandu), neonatal visits, iron-folic acid tablet (TTD) consumption by pregnant women, and living in poverty. The evaluation results show that GPR is superior with an AIC value of 505.68 and a Pearson Pseudo-R² of 0.1784. Based on the modeling results, iron tablet consumption and the number of poor residents have a significant effect, while active Posyandu does not. Furthermore, the categorical variable for basic immunization and the variable for neonatal visits showed anomalous results and were therefore eliminated to maintain model stability. This study demonstrates that GPR provides more reliable estimates to support targeted health intervention and poverty alleviation policies in Central Java.]]>
Copyrights © 2026
