Pemodelan Angka Kematian Bayi di Jawa Barat Menggunakan Pendekatan Analisis Regresi Spline dan Kernel

Abstract

The Infant Mortality Rate (IMR) is a very sensitive indicator of health service efforts, especially those related to newborns. IMR is also one of the problems that need to solve and the target of the SDGs number 3 (Good health and well-being). Java Province consists of 27 regencies/cities with an IMR of 3,26/1000 live births in 2019. The pattern of IMR data in West Java province had a pattern that changes at certain points so that the modeling is carried out using nonparametric regression. The selected nonparametric regression approach was spline regression which able to adapt more effectively with the characteristics of the data and kernel regression is easy to implementation. The explanatory variables used are life expectancy, the percentage of poor people, the open unemployment rate and the average length of schooling. The best model given by spline regression at 3 knot and kernel regression with bandwidth 1.2; 1.2; 1.1; and 1. Based model evaluation, the spline regression model's performance is better than the kernel regression with MSE, RMSE, and MAPE values are 0.66; 0.81, and 18.54%