PEMODELAN GEOGRAPHICALLY WEIGHTED REGRESSION TERHADAP FAKTOR-FAKTOR YANG MEMPENGARUHI ANGKA PUTUS SEKOLAH MENENGAH KEJURUAN DI PROVINSI SUMATERA UTARA

The proportion of kids who are of school age but are no longer enrolled or did not complete their education at a certain level is known as the dropout rate. The majority of dropouts are from vocational high schools. One of the reasons why students leave school is because the causes of dropouts are not accurately identified. This issue persists in the field of education. One issue with geographic heterogeneity is dropout. the development of geographical effects or spatial heterogeneity as a result of variations in each region's features and the connection between their distances. Geographically Weighted Regression (GWR) is one technique for analyzing spatially heterogeneous issues. The fixed kernel's weighting function and the adaptive kernel's weighting function in this research are both gaussian. The goal of this research was to choose the most appropriate model to utilize for the GWR model on the variables influencing the dropout rate for vocational high schools in North Sumatra Province. For each North Sumatra district or city, a distinct model is generated by this study. As compared to the multiple linear regression model with Ordinary Least Square (OLS) and the GWR model with fixed kernel weighting function gaussian, the GWR model with the adaptive weighting function of the gaussian kernel is the best model used to model the factors that influence the dropout rate for vocational high schools in North Sumatra Province. This is because it has the smallest AIC value of 321.7397 and the highest of 0.9756.