ESTIMATED MODEL OF URBAN LAND VALUE TOWARDS SINGLE VALUE FOR MULTIPURPOSE

• Main Author: Trianawati Sugito, Nanin
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/52906
• Institution: Institut Teknologi Bandung

The urban land value estimation model is a mathematical model to estimate the value of land in urban areas. Until now, there is no accurate land value estimation model that can be used as a reference for various purposes. A single value for multiple purposes is known as a single value for multipurpose, which is a single reference value that can be used as a reference for various needs. Land assessments still found problems. In conducting land appraisal, the number of tax objects is very large, so it requires hard work and very expensive costs. In addition, the results of land appraisal are inaccurate, as evidenced by the Sales Value of Tax Objects which does not reflect the actual value in the field. For this reason, it is necessary to refine a more comprehensive land value estimation model, using variables that are easy to calculate and determine for their characteristics to have higher accuracy. It is hoped that this land value estimation model will be easier for the appraiser or the community to understand. The research data uses the Average Indication Value data from the Regional Revenue Management Agency of Bandung City in 2007; data on transportation infrastructure and public infrastructure for Bandung City obtained from the Bandung City Spatial Planning Office in 2010; and Bandung City Spatial Planning 2004-2013 data from the Regional Development Planning Agency of Bandung City. The sample data is taken from the Average Indication Value data which has been overlaid with the variable determining the value of land and the variable of city service centers. Average Indication Value as land value data attached to land parcels in the form of polygons is then converted to centroids, which are the center points of the land parcels. There are 19 determinants of land value that affect land value in this study, namely roads, hotels, higher education, hospitals, public cemeteries, places of worship, schools, trade centers, sports stadiums, restaurants, industrial centers, terminals / stations, prisons, toll roads, landfills, localization / prostitution areas, airports, Base Transceiver Station (BTS) towers, and rivers. The city service center variable consists of 8 variables which are aspects of spatial structure that affect the modeling of land value estimates. The primary centers of Bandung are Alun-alun and Gedebage, while the secondary centers are Setrasari, Sadang Serang, Kopo Kencana, Turangga, Arcamanik, and Margasari. Each variable represents 10 centroid samples, so that in the development area coverage there are 270 samples. Estimation of land values is carried out using geostatistical methods, multiple linear regression, and non-linear regression. Model validation is used to determine the level of accuracy of the land value estimation model by calculating the standard deviation of the land value estimation model based on geostatistical methods, multiple linear regression, and non-linear regression. The land value estimation model based on geostatistical analysis, multiple linear regression, and non-linear regression was tested on the check point data. Check point data is the distribution of points in the coverage of the development area. The check point data selection is randomly done based on the determining variable of land value and the variable of city service centers. Check point data is used to test land estimation models based on geostatistical analysis, multiple linear regression, and non-linear regression. After using the check point, the amount of Root Mean Square Error (RMSE) is obtained from the geostatistical model, multiple linear regression, and non-linear regression. The multiple linear regression model is the best land estimation model for the western part of Bandung which includes the development areas of Bojonagara, Cibeunying, Karees, and Tegallega. The non-linear regression model is the best land value estimation model for the eastern part of Bandung which includes Gedebage and Ujungberung development areas. The land value estimation model based on geostatistical analysis produces the highest RMSE because the modeling only focuses on spatial variables and has not included the determinants of land values and city service centers. The land value estimation model that is close to the actual land value can be seen from the lowest RMSE model value so that it can be used as an accurate land value reference model. The multiple linear regression model is the best land value estimation model for the western part of Bandung which can increase the accuracy rate by 7.27% compared to the non-linear regression model. The non-linear regression model is the best land value estimation model for the eastern part of Bandung which can increase the accuracy rate by 14.26% compared to the multiple linear regression model. The novelty in this research is the integration of spatial variables, spatial variables (city service centers), and variables that determine land value to build land value estimation models in urban areas. Land value estimation models contribute to accelerating land valuation; field survey cost efficiency in data collection; This land value estimation model is expected to be a single value for various purposes; and can be used as a reference for urban areas in Indonesia.