IDENTIFIKASI BANDWIDTH NIR BERDASARKAN METODE GAUSS – GWR DENGAN INTERPOLASI TIN (STUDI KASUS: BANDUNG TIMUR)

• Main Author: Wahyu Ardhiwibowo, Fadhly
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/46954
• Institution: Institut Teknologi Bandung

The city of Bandung experienced a fairly rapid development, especially in East Bandung. These developments will have an impact on increasing land requirements, which will cause land values in the East Bandung region to increase. But in gathering information on the value of land in the field it takes a long time and costs are quite expensive. Therefore, in this study the Average Indication Value (NIR) in the Land Value Zone (ZNT) in East Bandung will be determined with a mathematical model that is weighting Geographically Weighted Regression (GWR) with Triangulated Irregular Network (TIN) interpolation. Areas that describe land values that are relatively the same as imaginary or real boundaries according to land use are called ZNT. Whereas the NIR is a mass assessment conducted in the administrative area of the village government which later produces a reference value that will serve as the basis for the NJOP determination. Geographically Weighted Regression is a spatial method that uses geographical factors as independent variables that can affect the dependent variable (Fotheringham et al, 2002). The optimum bandwidth of the GWR results at some sample points changes when TIN is rasterized. This will cause when testing with the test point, the resulting RMSE value will be influenced from the sample point where the optimum bandwidth value at some sample points changes during TIN interpolation. The reliability of modeling land prices with GWR based on TIN interpolation is 305,655.