Spatial Modeling of Food Security Index in Central Java Using Mixed Geographically Weighted Regression

Chamidah, Nur (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Saifudin, Toha (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Mahadesyawardani, Arinda (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Fauziah, Nathania (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Rahayu, Rizky Dwi Kurnia (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Siagian, Kimberly Maserati (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)
Wieldyanisa, Ezha Easyfa (Department of Mathematics, Faculty of Science and Technology, Universitas Airlangga)

Zero : Jurnal Sains, Matematika, dan Terapan
Vol 9, No 1 (2025): Zero: Jurnal Sains Matematika dan Terapan

Central Java plays an important role in Indonesia’s food security, ranking second nationally in the 2023 Food Security Index (FSI). However, nearly 45% of its districts/cities fall below the provincial average, reflecting spatial disparities. This study applies the Mixed Geographically Weighted Regression (MGWR) method to model the factors influencing FSI in Central Java by considering global and local spatial heterogeneity. Six clusters were formed based on similar characteristics. The MGWR model identifies that the factor of households not having access to clean water has a global negative effect which contributes 0.1710 points in decreasing the FSI, while population density is the dominant local factor that has a significant negative effect on the FSI in 21 districts/cities, covering approximately 60% of the area in Central Java. The MGWR model using a fixed Gaussian kernel outperforms global regression and GWR, with the lowest AIC, highest (93.11%), and a MAPE of 1.00838%.