Geographically weighted panel regression with adaptive bisquare kernel on stunting data in Southeast Sulawesi Province

Abstract

Stunting is a public health problem that shows spatial and temporal variations between regions. Differences in health, nutrition, environment, and human development characteristics cause the relationship between determinants and the number of stunting cases to be not homogeneous. This study aims to model the number of stunting cases in Southeast Sulawesi Province using Geographically Weighted Panel Regression (GWPR) with the weighting function of Adaptive Bisquare Kernel. The data used is secondary panel data for 17 districts/cities during the 2019–2023 period with the number of stunting cases as dependent variables and six independent variables representing health, nutrition, environment and human development factors. The results show that the GWPR model with the Adaptive Bisquare Kernel is able to accommodate spatial and temporal heterogeneity through estimating different local parameters in each district/city. The difference in local coefficients indicates that the influence of the determinants of stunting varies between regions. Thus, the GWPR approach provides more specific information about the characteristics of each region so that it can support the formulation of stunting control policies that are more targeted.