Application of Decision Tree Model in Mass Land Valuation: A Case Study in Vung Tau City

Abstract

Land valuation is one of many important tasks in land use and management. In order to be objective and scientific in land valuation, it is necessary to identify factors affecting land price and quantify the relationship between land feature factors and land value. The goal of the study is to apply the decision-tree regression model to build mass land pricing models in Vung Tau city. The decision-tree regression model is applied in land valuation by determining the relationship between a combination of land feature factors and corresponding averaged land price. Research on building the land pricing model with independent variables includes business density, area, road types, business activities and road width based on 883 survey samples. The final result of this tree model was categorized into 51 divided groups, and pruned trees included 15 divided groups. The importance of the independent variables, namely, business density, region, road level, business activities and road width are 100%, 83.48%, 78.87%, 58.78% and 10.03% respectively. The interpretation rate of the independent variables to the formation of land price model is 86%. The decision-tree regression model suggests another approach to land valuation theory.