Comparison of Negative Binomial Regression Model and Geographically Weighted Poisson Regression on Infant Mortality Rate in South Sulawesi Province

The number of infant mortality cases is an important indicator to assess the quality a country's public health. A studies argue that case has close relation living area condition and social status parents. Indirectly, life babies in country will impact nation's general. Therefore, many efforts are required reduce Indonesia. One steps could be done overcome this issue analyze causative factors. statistical method been developed for data analysis taking into account current spatial factors Geographically Weighted Poisson Regression (GWPR) with weighted Bisquare kernel function. Based on partial estimation GWPR model, there seven groups based significant variables affect deaths South Sulawesi Province. Of formed, first group Selayar Islands where all have effect. This needs concern provincial government improve facilities infrastructure Islands, course location which very far from city center can access drug reception, medical personnel so on. results Province using negative binomial regression approach bisquare weighting, it concluded model used best analyzing because AIC value. smallest 167.668.