Title: Application of Truncated Spline Nonparametric Regression For Modeling Income Inequality in Yogyakarta: A Panel Data Approach
Authors: Nashwa Carista, Toha Saifudin, Dita Amelia, M. Fariz Fadillah Mardianto
Volume: 9
Issue: 12
Pages: 80-88
Publication Date: 2025/12/28
Abstract:
Income inequality remains a persistent development issue in Indonesia, including Yogyakarta, and is commonly assessed using the Gini Ratio as key indicator. This research aims to model income inequality in Yogyakarta by analyzing the Gini Ratio using truncated spline nonparametric regression within a panel data framework. The analysis utilizes secondary data from the Central Bureau of Statistics for the 2020-2025 period, incorporating several socio-economic indicators as predictor variables. The truncated spline method is selected for its ability to accommodate both linear and nonlinear patterns without imposing strict functional assumptions. Model selection relies on the Generalized Cross Validation (GCV) criterion, with the optimal model obtained using three knots. The best model yields GCV value of 0.00045 and R-Square value of 99.53%, indicating an excellent fit. The findings show that key economic indicators significantly explain variations in Gini Ratio, demonstrating the effectiveness of nonparametric methods in capturing the complex structure of income distribution. These insights support evidence-based policymaking aligned with Sustainable Development Goals (SDGs) 10, which emphasizes reducing inequality within regions. Despite limitations related to aggregated panel data and the short observation period, this research provides methodological value by offering a flexible analytical framework that can complement traditional parametric approaches.