Analisis Regresi Logistik Spasial untuk Menduga Status Kemiskinan Desa di Kabupaten Majalengka

Abstract

Rural poverty status is influenced by several factors that are correlated between one rural area and the other. Logistic regression modeling with spatial weights is needed to determine the factors that affect rural poverty hotspots that are spatially correlated. Categorization of rural areas into hotspot (1) and non hotspot (0) was based on a hotspot Upper Level Set (ULS) Scan Statistic method which obtained 187 hotspot rural areas and 149 rural non hotspot. Rural hotspot areas have characteristics similar to non rural hotspot areas in various aspects. Logistic regression models with Spatial variable enhances logistic regression model without Spatial variable in predicting the status of rural poverty areas in Majalengka. Logistic regression models with Spatial variable has a higher correct classification rate (CCR) compared to logistic regression without Spatial variable. The explanatory variables that have a significant influence towards the status of rural poverty are Percentage of Farm Worker Families variable and Spatial variable.