基于GWR的住宅地价相对修正方法研究——以深圳市为例

研究目的:提出一种适用于地产市场活跃地区的住宅地价相对修正方法,解决传统市场比较法过分依赖估价师经验的问题。研究方法:运用GWR模型,量化影响因素对住宅地价作用的空间分异规律,确定影响因素单位变化对应的地价修正百分比;运用空间单元相似度评价方法,自下而上生成住宅地价修正分区。研究结果:从GWR模型原理出发,建立了基于客观数据的修正公式;构建了从回归模型构建到修正分区建立,再到地价修正的完整技术路线;研究提出的修正方法经真实数据检验取得了较好的效果。研究结论:本文构建的住宅地价相对修正方法在估价工作中具有实操可行性;地价修正精度与GWR模型构建精度正相关;该方法更加适用于具备高密度地价样本点的地产市场活跃地区。

Study on the Correcting Method of Residential Land Price based on GWR: A Case Study of Shenzhen City

The purpose of the study is to propose a correcting method of residential land price, which is suitable for the regions with active estate market, to solve the problem of traditional market comparison approach that relies too much on experience of the appraisers. The research methods are as follows. The GWR model is introduced to measure the spatial differentiated law of the impact factors of residential land price, and to determine the correcting percentage in light of each unit change of impact factors. A spatial unit similarity evaluation method is developed to generate the residential land price correcting zone from bottom to top. The research results show that starting from the principle of GWR model, the correcting formula based on objective data is established and a complete technical route containing the regression model building, correcting zone generating and land price correcting is developed. The proposed method achieves good effect in the test by using real data. In conclusion, the proposed method has feasibility in the actual land appraisal work; the precision of land price correcting is positively associated with the precision of GWR model; and this method is more suitable for the regions with active estate market which have high density of land price sample points.