Perbandingan Regresi Nonparametrik Kernel dan Spline pada Pemodelan Hubungan antara Rata-Rata Lama Sekolah dan Pengeluaran per Kapita di Indonesia

Abstract

Poverty remains a major issue in developing countries, including Indonesia. In 2021, Indonesia’s poverty rate reached 10.14%, or approximately 27.5 million people (BPS). Poverty alleviation is a primary goal within the Sustainable Development Goals (SDGs). Two important indicators for measuring poverty are per capita expenditure and average years of schooling, which can aid in formulating policies to reduce poverty. This study analyzes the relationship between average years of schooling and per capita expenditure in 2023 using nonparametric regression methods, specifically kernel and spline regression. The kernel regression analysis yielded an optimal bandwidth of 0.860 and a minimum GCV of 0.574. However, the truncated spline method, with one optimal knot, a minimum GCV of 0.5263514 at the 3rd order, and the smallest MSE of 0.4097892, proved to be more accurate in describing the relationship between the two variables. The study concludes that the truncated spline method is superior in modeling the relationship between per capita expenditure and average years of schooling, providing valuable insights for policy formulation aimed at poverty alleviation in Indonesia.