Analysis of special index of stunting treatment through confidence interval of longitudinal semiparametric spline model parameters

Abstract

Stunting is a major concern in many nations, including Indonesia, and it has long-term implications for human resource quality. The Special Index of Stunting Treatment (SIST) is one of the social population problems that can be modeled with regression. The combination of parametric and nonparametric regression techniques is known as a semiparametric regression. This study used longitudinal data, obtained from repeated observations within a certain time span on the same individual in sequence. This study's estimation method makes use of a truncated spline, which has great flexibility and statistical interpretation. In the statistical inference of the parameter, the calculation of the confidence interval is crucial. Providing an estimated range for a population parameter based on two boundary points enables quantification of the level of confidence in the accuracy and reliability of the estimate. Six dimensions are formed in the acceleration of stunting reduction: health, education, nutrition, housing, food, and social protection. Results of the confidence interval estimation of truncated semiparametric spline regression parameters on longitudinal data. The optimal model was obtained using three knot points, resulting in a minimum GCV of 12.7066, an MSE of 2.0225, and R² 93.66%. This study shows that this model is feasible and the predictor variables have a negative and positive influence on the Special Index of Stunting Treatment (SIST).