Geographically weighted regression in malnourished toddlers with adaptive kernel bi-square weighting

Spatial data modeling cannot use linear regression due to the presence of unfulfilled assumptions, namely heteroscedasticity. One of the modeling that can be done is Geographically Weighted Regression (GWR) with weighting, one of which is Adaptive Kernel Bi-square. This study aims to prove that the GWR Adaptive Kernel Bi-square model can analyze and interpret factors that have a significant effect on cases of malnourished toddlers in East Java Province. This study used GWR Adaptive Kernel Bi-square analysis using secondary data on the Health Profile of East Java Province in 38 districts/cities in 2018. The results of GWR modeling with the Adaptive Kernel Bi-square weighting function, namely the value of the coefficient in each region is different and the difference in predictor variables has a significant effect. The GWR Adaptive Kernel Bi-square model can analyze and interpret factors that have a significant effect on cases of malnourished toddlers in East Java Province, so this study can be a reference in analyzing influential factors.