<![CDATA[Regression methods that take into account spatial aspects are often influenced by the geographical conditions of each observation location. In this study, a regression model was constructed to predict the relationship between the value of the illiteracy rate and several independent variables using Multiscale Geographically Weighted Regression (MGWR) in South Sumatra province. These independent variables are the population (X1), the pure elementary school participation rate (X2), the pure junior high school participation rate (X3), the number of elementary school teaching staff (X4), the number of junior high school teaching staff (X5), and the percentage of poor people (X6). MGWR was chosen because of the use of bandwidth in each variable, so it is expected to provide a model accuracy that is thought to be more accurate to the data. One of the ABH models found in Palembang City is Y= -0.0187+0.55X1-0.1748X2 - 0.0062X3 - 1.6129X4 + 0.5394X6.Keywords: Bandwidth, GWR, Illiteracy Rate, MGWR, Spatial Heterogenity. AbstrakMetode regresi dengan memperhatikan aspek spasial sering kali dipengaruhi oleh kondisi geografis dari masing-masing lokasi pengamatan. Pada penelitian ini dikontruksi model regresi untuk memprediksi hubungan nilai Angka Buta Huruf (ABH) dengan beberapa variabel bebas menggunakan Multiscale Geographically Weighted Regression (MGWR) di provinsi Sumatera Selatan. Variabel-variabel bebas yang dimaksud adalah jumlah penduduk (X1), angka partisipasi murni SD (X2), angka partisipasi murni SMP (X3), banyak tenaga pendidik SD (X4), banyak tenaga pendidik SMP (X5) , dan persentase penduduk miskin (X6). MGWR dipilih karena memungkinkan penggunaan bandwidth pada setiap variabel, sehingga diharapkan mampu memberikan suatu ketepatan model yang diduga lebih akurat terhadap suatu data. Salah satu model ABH yang terdapat di Kota Palembang adalah Y= -0.0187+0.55X1-0.1748X2 - 0.0062X3 - 1.6129X4 + 0.5394X6.]]>
