PEMODELAN ANGKA HARAPAN HIDUP LAKI-LAKI DAN PEREMPUAN DI INDONESIA MENGGUNAKAN PENDEKATAN REGRESI SEMIPARAMETRIK SPLINE TRUNCATED

Abstract

Life expectancy (LE) ) is an indicator of health status used to improve the quality of human life. This study models the LE of men and women to determine the factors that have a significant effect using a truncated spline semiparametric regression approach. This regression is done by looking at the optimum knot point from the minimum Cross Validation (CV) value. Based on the method used, the best model produced is with three knot points. The results of the analysis on male LE are that all independent variables have a significant effect, with a minimum CV value of 5.30 and a coefficient of determination ( ) of 96.70%. Meanwhile, in female LE, there are three independent variables that have a significant effect out of a total of eight independent variables, with a minimum CV value of 11.05 and a coefficient of determination ( ) of 86,82%.