摘要
从空间自相关的角度出发,在房地产估价市场比较法的基础上,应用商业类房地产数据,将经典线性回归模型、空间滞后模型和空间误差模型进行比较研究。在误差分析中,空间误差模型得出的评估值与真实值误差率在正负20%之间的样本数占到样本总数的69.64%,高于经典线性回归模型的56.25%和空间滞后模型的44.2%,由此说明该组房地产数据中存在空间自相关。用空间误差模型可以提高批量评估结果的精确度。
From the perspective of spatial autocorrelation,based on the valuation of real estate market comparison method,utilizes commercial property data from Shenzhen to investigate OLS model,Spatial Lag Model(SLM)with SEM. Then,conducts an error analysis,the sample number with error rate of assessed value and real value between plus or minus 20% accounted for 69.64% of the total number of samples by SEM,demonstrates SEM performs better than OLS Model and SLM with 56.25 percent and 44.2 percent,the result shows the spatial autocorrelation in the group of real estate date.
出处
《建筑经济》
2015年第2期62-67,共6页
Construction Economy
