A multi-predictor nonparametric local linear generalized Poisson regression model for overdispersed count data

Generalized Poisson regression is widely used for modeling overdispersed count data; however, existing nonparametric local linear generalized Poisson regression methods are generally restricted to a single predictor, limiting their applicability to datasets involving multiple predictors. This study extends the nonparametric local linear generalized Poisson regression framework to accommodate multiple predictors, namely multi-predictor local linear generalized Poisson regression (MP-LLGPR). The proposed method provides a flexible approach for modeling overdispersed count response variable with locally varying relationships involving multiple predictors. The method is empirically validated in a healthcare application using stunting count data from 44 regencies and cities in Kalimantan, Indonesia. The empirical validation demonstrates the applicability of the proposed method for modeling overdispersed count data with multiple predictors and provides a comparison with the corresponding global generalized Poisson regression model.