基于GWR模型的武汉市住宅地价空间分异及影响因素研究

发布时间：2018-04-17 08:16

本文选题：住宅地价 + GWR模型　；参考：《华中农业大学》2014年硕士论文

【摘要】：土地价格是土地经济作用的反映,受土地市场上供求关系的影响,土地价格始终处于变动状态。从空间的层面来看,地价及其变动具有明显的区域与区位差异,即地价的空间分异特征。研究地价空间分异特征的形成、影响因素及其规律性,有助于充分发挥价格机制对于土地市场的调控作用,加强地价管理和调整土地的供求区位,从而在城市“精明增长”理念的指导下,落实土地利用的“精明规划”,促进城市土地资源的效率配置。以往同类研究主要基于线性回归或特征价格模型展开,此类模型基于城市空间匀质的假设前提,认为在城市内部的不同区位上,住宅地价的各影响因素对住宅地价的贡献度均等。然而由于城市区域内部自然及社会经济资源分布的不均等性,在不同的微观区位上,住宅地价的影响因素对于住宅地价的边际增长作用应该是可变的。地理加权回归模型(GWR)扩展了传统的回归模型框架,通过在线性回归模型中假定回归系数是观测点地理位置的函数,从而将数据的空间特性纳入到模型中,为分析回归关系的空间特征创造了条件,并通过优化权重,使得模型能够对空间非稳定性数据进行局部估计。因此,相对于传统的回归模型,使用GWR模型来研究城市地价的空间分异规律,可以更为深入的研究城市地价空间的形成机理及影响因素的作用机制。本文选择城市住宅地价为研究对象,以武汉市主城区为研究区域,分析了2003-2012年期间武汉市的住宅地价的空间分异特征,发现武汉市的住宅地价总体上仍表现为单中心结构,并以该中心为核心向外圈层式递减；在此基础上,选取地块面积、容积率、大学、中小学、医院、主干道、地铁、生活品市场、水域及公园10个住宅地价微观影响因素,构建GWR模型分析住宅地价空间分异特征的形成规律,结果表明：容积率、主干道、生活品市场及教育资源对住宅地价贡献度的区位差异显著。一方面表现出城市居民在选择居住区位时对生活便利程度、教育资源及交通条件的偏好,另一方面也说明了教育资源、生活品市场以及交通条件等因素在研究区内的分布不够均衡。相对于传统的线性回归或特征价格模型,GWR模型更适合于从微观区位的角度为住宅空间规划和地价更新提供依据,更为符合城市精明增长的规划理念,为优化配置城市土地资源提供理论实践指导。
[Abstract]:The land price is the reflection of the land economy, and the land price is always in the state of change under the influence of the supply and demand in the land market.From the spatial aspect, the land price and its change have obvious regional and regional differences, that is, the spatial differentiation of land price.The study of the formation, influencing factors and regularity of spatial differentiation of land price is helpful to give full play to the role of price mechanism in regulating the land market, to strengthen the land price management and to adjust the land supply and demand location.Therefore, under the guidance of the urban "smart growth" concept, the "smart planning" of land use can be carried out to promote the efficient allocation of urban land resources.In the past, similar studies were mainly based on linear regression or characteristic price model. This kind of model was based on the assumption of urban spatial homogeneity.The contribution of all factors to residential land price is equal.However, due to the uneven distribution of natural and social economic resources in urban areas, the influence factors of residential land price on the marginal growth of residential land price should be variable in different micro-location.Geo-weighted regression model (GWR) extends the traditional regression model framework. By assuming that the regression coefficient is a function of the geographical location of the observation point in the linear regression model, the spatial characteristics of the data are incorporated into the model.Conditions are created for the analysis of spatial features of regression relations and local estimation of spatial instability data can be made by optimizing weights.Therefore, compared with the traditional regression model, the use of GWR model to study the spatial differentiation of urban land price, can be more in-depth study of the formation mechanism of urban land price space and the action mechanism of influencing factors.This paper chooses the urban residential land price as the research object, taking the main urban area of Wuhan as the research area, analyzes the spatial differentiation characteristics of the residential land price in Wuhan from 2003 to 2012, and finds that the residential land price in Wuhan is still characterized by a single central structure.On the basis of this, we select the microcosmic influencing factors of land area, volume ratio, university, primary and middle school, hospital, trunk road, subway, life product market, water area and park 10 residential land price.The GWR model is constructed to analyze the spatial differentiation characteristics of residential land price. The results show that: volume ratio, main road, market of living goods and educational resources have significant regional differences to the contribution of residential land price.On the one hand, it shows the preference of urban residents for living convenience, educational resources and traffic conditions when they choose residential areas. On the other hand, it also shows that educational resources.The distribution of market and traffic conditions in the study area is not balanced.Compared with the traditional linear regression model or the characteristic price model, GWR model is more suitable to provide the basis for housing spatial planning and land price renewal from the perspective of micro-location, and is more in line with the urban smart growth planning concept.To optimize the allocation of urban land resources to provide theoretical and practical guidance.
【学位授予单位】：华中农业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F301.4;F224

【参考文献】