Spatial Analysis of Food Security in Indonesia 2022 Using a Geographically Weighted Logistic Regression Model

Abstract

The crisis in food security is a problem in many countries including Indonesia. The crisis is important to pay attention to, in order to realize sustainable food security in accordance with the second Sustainable Development Goals (SDGs) of zero hunger. Achieving food security cannot be separated from making appropriate and targeted policies. Policy making must be supported by accurate, comprehensive and systematic food security information. This study aimed to classify districts/cities in Indonesia into food-secure and food-prone areas and analyze the factors that influence the food security status. The data were analyzed descriptively and inferentially where clustering with k-means method for food security status classification and Geographically Weighted Logistic Regression (GWLR) modeling with Adaptive Gaussian kernel function weights which provided the best model with the smallest Akaike Information Criterion corrected (AICc) compared to other kernels to see the spatial influence in determining the food security condition of a district/city. The results showed that in 2022 there were 388 districts/cities with food security status and 126 districts/cities with food insecurity status. In addition, of the 9 variables used in clustering, there are 5 variables used in GWLR modeling. Three of them are significant in all regions in Indonesia, while 1 variable is locally significant in some regions.