【作者】 曾晖；

【导师】 杨平；

【作者基本信息】 南京林业大学 ， 森林工程， 2012， 博士

【摘要】 住宅价格问题目前已成为我国社会各阶层普遍关注的问题，中国的住宅价格不仅仅是经济问题，而且涉及到政治问题。因此，对住宅价格进行理性的、实事求是的分析和研究具有非常重要的意义。本文以住宅特征价格理论为基础，考虑区位因素和一定期限的时间因素，通过全面考察住宅价格的影响因素，揭示住宅价格在一个有限区域内可能存在的分布规律及随时间变化的演化趋势。实证研究选取南京市（代表大城市）和江门市（代表中小城市）具有代表性的住宅小区为研究对象，收集、处理各小区均价以及影响均价的特征因素，形成了两个城市住宅小区均价数据库和影响特征因素数据库，对两个城市住宅小区均价进行了探索性空间数据分析和确认性空间数据分析，提出城市住宅小区均价存在着全局空间相关性和局部空间相关性，并将城市的住宅小区根据局部空间相关性的类型进行了分类；通过应用空间滞后模型和空间误差模型对住宅小区均价进行回归分析，验证了模型的合理性和适用性;应用地理加权回归模型分析住宅价格影响因素的空间非平稳性；将空间计量经济分析结果与地统计学中的插值技术相结合，应用泛克里格插值法对两个城市的住宅小区均价进行空间插值，揭示了两个城市住宅小区均价在一段时间内所具有的空间分布规律及变化趋势，即时空分布规律。最后，为缓解住宅小区均价峰值中心的压力，使城市住宅价格具有合理的分布模式，根据两个城市住宅小区均价的空间分布和时空演化规律，提出了具有实践意义的建议。更多还原

【Abstract】 Housing prices has drawn universal concern of all walks of life in China. The issue ofhousing prices is not only an economic issue but also a political one. Subsequently, a rational andrealistic analysis and study of housing prices is of vital signiifcance.Based on hedonic housing price theory, taking location factor and time factor within a peirodinto consideration, and via comprehensive study of the influential factors on housing pirces，thepaper reveals possible distirbution rules and evolution trend over time of housing prices. Theresidential communities in Nanjing (representative of big cities) and Jiangmen (representative ofsmall and medium cities) are taken as the empirical study. Via collection and disposal of theaverage prices of the residential communities and the influential factors of the average pirces，average price data base and influential factor data base of the residential communities in the twocities take shape. Through exploratory spatial data analysis and conifrmative spatial data analysis,it is concluded that the residential communities in the two cities are of overall spatial correlationand local spatial correlation. Besides, the residential communities are classified according tolocal spatial correlation. The application of spatial lag model and spatial error model onregression analysis of the average prices of the residential communities veriifes the rationalityand applicability of the foregoing models. Geographic weighted regression model is applied toanalyze spatial instability of the influential factors of the residential communities. By combiningspatial econometric analysis result and interpolation technology in geostatistics, and usingUniversal Kriging method to carry out spatial interpolation of the residential communities, thespatial distirbution rules and evolution trend within a peirod of time, i.e. spatial and temporaldistirbution rules are revealed.Finally, in a bid to ease the pressure of peak center of the residential communities, and torationalize the distirbution mode of urban housing pirces，practical suggestions are proposedbased on the spatial distirbution and spatial and temporal evolution rules of the average prices ofthe residential communities in the two cities.更多还原