#TITLE_ALTERNATIVE#

• Main Author: SUSANTO (NIM 15105056); Pembimbing : Ir. Albertus Deliar, MT. dan Dr. Ir. D.Muhally Hakim, M., HENDRA
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/13520
• Institution: Institut Teknologi Bandung

Main purpose of this research is to investigate Geographically Weighted Regression (GWR) method compared to linear regression method for land value modeling based on Geographic Information System. GWR method or regression with geographical weight has advantage such as ability to include local criteria in modeling calculation (Fotheringham et al., 2002). This advantage probably very compatible to applied in land value modeling, especially because land value very influenced by location factor (Atack, 1998) ; (Hariadi, 2003).<p> With local criteria represented by weight for each value sample in land value modeling, this research gets result that land value produced by GWR method are better compared to land value produced by linear regression method. This proven by correlation coefficient for GWR method modeling has value of 0.320 and Root Mean Square error 822535,393 whereas with linear regression, correlation coefficient only reaches 0.277 and Root Mean Square error 933094,074.