ESTIMATING GEOGRAPHICALLY WEIGHTED NEGATIVE BINOMIAL REGRESSION (GWNBR) MODEL AND MAPPING SPATIAL PATTERNS ON ANEMIA DATA IN INDONESIA

ABSTRACT: Geographically Weighted Negative Binomial Regression (GWNBR) is a spatial regression approach designed to account for spatial heterogeneity by incorporating geographic coordinates (latitude and longitude) as spatial weights while simultaneously addressing overdispersion in count data. The empirical application in this study utilized data on anemia cases among women of childbearing age (WCA) across 33 provinces in Indonesia. The GWNBR model was employed to estimate local regression parameters and examine the spatial variation in regression coefficients. The analysis revealed that at least one explanatory variable significantly influenced the response variable in each province. In the central and eastern regions of Indonesia, the number of anemia cases per 100 WCA increased proportionally with increases in the number of WCA affected by pneumonia, malaria, and hepatitis. In contrast, western Indonesia exhibited distinct spatial patterns, where anemia cases per 100 WCA were positively associated with the number of WCA residing in rural areas (particularly in Java and Kalimantan), the number of WCA affected by acute respiratory infections (ARI) in Sumatra and Kalimantan, and the number of WCA diagnosed with tuberculosis. Keywords: Negative binomial regression, Geographically Weighted Negative Binomial Regression (GWNBR), anemia, Women of Childbearing Age (WCA