Understanding the Effects of Influential Factors on Housing Prices by Combining Extreme Gradient Boosting and a Hedonic Price Model (XGBoost-HPM)

The characteristics of housing and location conditions are the main drivers spatial differences in prices, which is a topic attracting high interest both real estate geography research. One most popular models, hedonic price model (HPM), has limitations identifying nonlinear relationships distinguishing importance influential factors. Therefore, extreme gradient boosting (XGBoost), machine learning technology, HPM were combined to analyse comprehensive effects factors on prices. XGBoost was employed identify order adopted reveal value original non-market priced results showed that combining two models can lead good performance increase understanding variations Our work found (1) five important variables for Shenzhen prices distance city centre, green view index, population density, property management fee economic level; (2) space quality at human scale had prices; (3) some traditional factors, especially related education, should be modified according development market. demonstrated multisource geo-tagged data fusion framework, integrated HPM, practical supports between findings this article provide essential implications informing equitable policies designing liveable neighbourhoods.