<![CDATA[Poverty in Central Java Province remains a significant multidimensional issue, characterized by socio-economic disparities across regencies and municipalities and high rates of school dropouts among children. This study aims to evaluate the influence of socio-economic variables on poverty depth (P1 Index) at both local and global levels. The approach employed is the Geographically Weighted Logistic Regression Semiparametric (GWLRS), which integrates local and global effects. The model uses two types of spatial weights: Adaptive Gaussian Kernel and Queen Contiguity, with predictor variables including Dependency Ratio, Minimum Regional Wage (UMK), Number of Industries, Open Unemployment Rate (TPT), Adequate Housing, and Sanitation. Parameter estimation was conducted using Maximum Likelihood. The results indicate that the Dependency Ratio, Adequate Housing, and Sanitation are locally significant in 2–5 regions. Local coefficients for the Dependency Ratio range from 0.42 to 11.57 (mean 3.33), Adequate Housing from -15.808 to -2.371 (mean-5.95), and Sanitation from 1.86 to 17.27 (mean 5.71). The model correctly predicts 31 out of 35 cases, yielding an accuracy of 91.4%. The Number of Industries, UMK, and TPT are not globally significant, indicating that their effects are more stable across regions. In conclusion, the GWLRS model effectively captures the spatial heterogeneity of poverty determinants and provides quantitative insights that can support more targeted, location-based poverty alleviation policies in Central Java Province.]]>
Copyrights © 2026
