Analyzing factors affecting land prices in urbanized areas using machine learning: A basis for future 3D property valuations

Peyman Jafary, Davood Shojaei, Abbas Rajabifard, Tuan Ngo

Accurate land valuation promotes fairness and efficiency in the property market, aiding property development and land use planning. Land price is also a key component for property valuation using cost approach, particularly with 3D models like Building Information Modeling (BIM). Identifying and creating a database on value-related features is essential to establishing a robust land valuation system. Advanced Machine Learning techniques can handle non-linear relationships between land value and the driving factors. This study analyzes factors affecting land prices in Melbourne Metropolitan, considering a wide range of features. The importance of each factor is calculated through an ensemble method based on four techniques: Random Forest, XGBoost, Recursive Feature Elimination (RFE) and Mutual Information (MI). The results show that land area, longitude, land use, latitude, distance to the Central Business District (CBD), elevation, mortgage rate and primary school zone have the highest impact on the land value in the study area.

Event: World Bank Land Conference 2024 - Washington

Document type:Analyzing factors affecting land prices in urbanized areas using machine learning: A basis for future 3D property valuations (782 kB - pdf)