Analisis Regression Splines untuk Memodelkan Hubungan Angka Harapan Hidup dengan Akses Air Minum Layak di Indonesia

Spline Regression Analysis to Model the Relationship between Life Expectancy and Access to Safe Drinking Water in Indonesia

Abstract

Life Expectancy (AHH) is one of the main indicators in assessing the quality of life and human development of a region. One factor suspected of contributing to AHH is household access to safe drinking water, as it is closely related to disease prevention, nutritional status, and the overall quality of public health. The relationship between the two variables is suspected to be non-linear, so the parametric regression approach is considered inadequate to capture the true pattern of the relationship. This study aims to analyze the relationship between the percentage of access to safe drinking water and AHH in 34 Indonesian provinces in 2024 using a nonparametric regression approach, specifically spline regression. Secondary data were obtained from the Central Bureau of Statistics, with AHH as the dependent variable and the percentage of access to safe drinking water as the independent variable. The analysis was conducted using R software through the stages of descriptive statistics, data visualization, estimation of a simple linear regression model as a comparison, and spline regression modeling with the selection of optimum knot points based on the smallest Generalized Cross Validation value. The results showed that a spline regression model with nine knots for women's AHH and five knots for men's AHH produced an R² value of 0.3675, higher than the OLS model of 0.3447. The relationship between the two variables was positive but nonlinear, with increases in AHH tending to slow down at higher levels of access to improved drinking water. This finding emphasizes the importance of prioritizing policies to increase access to improved drinking water in areas with low access to provide a more significant health impact for the community.