Abstract

Geographically weighted regression (GWR) is one of model in spasial regression. It used when the spatial regression had heteroschedasticity. Each location will have different model based on the significant coeficient in every explanatory variable at each location. Then if the respon variable is binary data so the model will be mentioned as Geographically Weighted Binary Logistic Regression. The study has aim to applicate it in real data. The Data taken from PODES which consist of one respon variable with binary type and three explanatory variable with ratio type. The respon variable is poverty category for each municipality and the independet variable is percentage of trade village, percentage of service village and ratio of school pervillage. This research gives Global Model and Local Model. At Global Model only two explanatory variables which has significant influence to the poverty category with AIC value about 122,563. At Local Model gives different result based on the significant coefficient in each municipality.

Authors and Affiliations