Locally kernel weighted maximum likelihood estimator for local linear multi-predictor poisson regression

This is the authors' abstract. We don't add key points for this paper.

We introduce a new multi-predictor regression model based on the Poisson distribution using a local linear approach called the local linear multi-predictor Poisson regression. The optimal bandwidth in this study was selected based on the maximum likelihood cross-validation (MLCV) value. The locally kernel-weighted maximum likelihood estimator is used to estimate the regression curve at a given point. Parameter estimation was performed using the Newton-Raphson iteration method. The superior points in this research are:•A new model in regression to model multi-predictor case Poisson regression problems using a local liner approach•Optimal bandwidth selection using MCLV•Application of multi predictor case Poisson regression problems using a local liner approach to health data; namely the stunting case in East Kalimantan.