应用改进的时空地理加权模型分析城市住宅价格变化

Analysis of Urban House Price Variations by Improved Geographically and Temporally Weighted Model

以深圳市为例对城市住宅价格进行时空地理加权统计回归分析,揭示城市住宅价格在时间与空间方面的变化趋势。研究结果表明,时空地理加权回归模型不仅能够分析相关因素在空间的差异性上对住宅价格具有重要影响,还能够揭示其在时间上的差异性,得出相关因素对住宅价格影响是随时间和空间变化的特点。依据模型的拟合度标准,GTWR能够从OLS的0.617,GWR的0.736提高到0.895。AIC标准的统计信息则表明,GTWR模型明显优于OLS与GWR模型。

This study intends to investigate spatiotemporal non-stationarity of the housing market dynamics in the City of Shenzhen with the improved geographically and temporally weighted model（GTWR） for exploiting spatiotemporal variations on house price.Our results show that there have been substantial benefits in modeling spatial and temporal non-stationarity simultaneously along with the variations across the study area.The GTWR model can improve the goodness-of-fit of the OLS model and the GWR model from 0.617 and 0.736 to 0.895 in terms of R-square.The AIC test corroborated that the improvements made by GTWR over the GWR and OLS models were statistically significant.The study concludes that housing market dynamics might be better understood by the combined effects of housing temporal and spatial attributes.