Garis Kemiskinan di DKI Jakarta: Analisis Regresi Non-Linear dengan GAM

Abstract

This study analyzes the poverty line in DKI Jakarta using non-linear regression with the Generalized Additive Model (GAM) and considering spatial aspects through Eigenvector Filtering. Secondary data from BPS DKI Jakarta for the period 2018-2022 is used with independent variables in the form of Expenditure per Capita (PPK), Open Unemployment Rate (TPT), and Human Development Index (HDI). The results show that PPK has a significant non-linear relationship with the poverty line, while TPT and HDI have no significant effect. Spatial analysis reveals patterns of poverty distribution that reflect differences in economic structure and access to public services in different areas of Jakarta. These findings emphasize the need for poverty alleviation policies that focus on increasing people's purchasing power and equalizing economic access across administrative regions.