PEMODELAN FAKTOR-FAKTOR YANG BERPENGARUH TERHADAP ANGKA BUTA HURUF DI PROVINSI SULAWESI SELATAN DENGAN GEOGRAPHICALLY WEIGHTED LOGISTIC REGRESSION (GWLR)

Abstract

Geographically Weighted Logistic Regression (GWLR) is the development of a logistic regression model applied to spatial data from non-stationary processes with categorical response variables. The high rate of illiteracy is one of the crucial problems in the field of education that has not been resolved to date. South Sulawesi is the 4th province with the highest percentage of illiteracy in Indonesia in 2022. This research aims to obtain the GWLR model and the factors that have a significant influence on the illiteracy rate in South Sulawesi in 2022. In this research, we compare three functions Kernel weightings are Adaptive Gaussian Kernel, Adaptive Bisquare Kernel, and Adaptive Tricube Kernel. Selection of the best model uses the smallest AIC value. The results of this research are that the GWLR model with the Adaptive Tricube Kernel weighting function is the best model in modeling cases of illiteracy in South Sulawesi in 2022 which is obtained based on the smallest AIC value and the factor that has a significant influence on the illiteracy rate is the Open Unemployment Rate (X₁), percentage of poor population (X₂), Elementary School Enrollment Rate (X₃), and area with city status (X₄).