基于GWR模型的上饶市住宅地价空间分异研究

发布时间：2018-01-01 00:13

本文关键词：基于GWR模型的上饶市住宅地价空间分异研究　出处：《江西师范大学》2014年硕士论文　论文类型：学位论文

【摘要】：城市土地是城市居民生产生活的主要场所，也是经济、政治、文化、教育、科技和信息传播的中心。城市地价则直接反映了土地市场和土地供需状况的健康状况，更是政府用来进行宏观调控的重要工具。随着国民经济的快速发展，如何通过研究城市地价的空间分异情况及其影响因素的状况，为城市制定合理的供地计划及政策，实现城市土地资产的可持续利用和防止土地资产流失也逐渐成为了土地资源管理的重要课题。在国内外有关城市地价的空间分异及其影响因素的研究有很多，大部分的研究都是通过量化影响地价的因子变量，利用统计分析的方法建立特征价格模型来表现整个研究区域地价的空间变化情况。这样的地价空间结构分析往往只能反映出全域的地价空间变化状况，对于研究区域的地价局部空间变化则无法很好的诠释。而且从评价结果来看，特征价格模型结果单一且欠缺直观性。因此需要选择更优的方法帮助分析地价的空间分异，即地理加权回归（GWR）模型。本文在国内外地价空间分异研究现状和理论研究的基础上，对城市地价的内涵进行了界定，以上饶市中心城区为研究区域，通过对上饶市2009~2012年土地出让市场情况的数据调查和量化整理，首先利用探索性空间分析（ESDA）方法，探索了上饶市住宅地价分布的空间数据结构和空间趋势变化情况，结果表明上饶市住宅地价数据基本符合正态分布，且地价在经度方向和纬度方向均呈现倒U型，尤其是经度方向的地价曲线值由南至北呈现低—高—低的分布。然后采用Geoda软件对上饶市住宅地价进行空间自相关分析，并对地价数据进行克里金插值利用Arcgis工具做不同方向上饶市的住宅地价剖面图。结果表明上饶市住宅地价在空间上呈现一定的集聚效应，且上饶市住宅地价分布为“单核”空间结构，表现为由市中心向外围递减的趋势，另外区域内还存在地价的“冷点”和“热点”。接着，通过对权重函数的选择和带宽的确定采用了地理加权回归（GWR）模型对上饶市住宅地价，详细探讨了上饶市住宅地价与其影响因素之间的关系。通过方差分析、事件智能诊断和蒙特卡罗显著性检验等对GWR模型的结果进行比较和分析，，确立了调整型的空间核作为解释住宅地价空间结构的最终模型。并结合了GIS工具分析了各个影响因素对住宅地价的空间分布差异。结果表明医院、中学、小学、公园、容积率和绿化率等影响因子对住宅地价影响显著，且各个影响因素的影响程度随着空间位置的改变，对地价的影响程度也不同。最后，对两种模型方法的拟合优度R2、残差平方进行了结果分析，发现GWR模型都比特征价格模型具有更好的解释结果。
[Abstract]:Urban land is not only the main place for urban residents to produce and live, but also the economy, politics, culture and education. Urban land price directly reflects the health of the land market and land supply and demand, and is an important tool used by the government to carry out macro-control. With the rapid development of the national economy. How to make a reasonable land supply plan and policy by studying the spatial differentiation of urban land price and its influencing factors. Realizing the sustainable use of urban land assets and preventing the loss of land assets have gradually become an important subject of land resource management. There are many researches on the spatial differentiation of urban land price and its influencing factors at home and abroad, most of which are by quantifying the factor variables that affect the land price. The method of statistical analysis is used to set up the characteristic price model to express the spatial change of the land price in the whole study area, and the spatial structure analysis of the land price can only reflect the spatial change of the land price in the whole region. The local spatial change of the land price in the study area can not be well explained, and from the evaluation results. The results of the feature price model are simple and lack of intuition.Therefore, it is necessary to choose a better method to help analyze the spatial differentiation of land price, that is, the geographical weighted regression model (GWR). Based on the present situation and theoretical research of spatial differentiation of land price at home and abroad, this paper defines the connotation of urban land price, taking Shangrao central city as the research area. Through the data investigation and quantitative arrangement of the land transfer market in Shangrao City from 2009 to 2012, the exploratory spatial analysis (ESDA) method is first used. The spatial data structure and spatial trend of residential land price distribution in Shangrao are explored. The results show that the data of residential land price in Shangrao basically accord with normal distribution. And the land price in longitude direction and latitude direction are inverted U-shaped. Especially the value of land price curve in longitude direction from south to north presents a low-high-low distribution. Then the Geoda software is used to analyze the spatial autocorrelation of residential land price in Shangrao city. And carry on the Kriging interpolation to the land price data to use the Arcgis tool to make the residential land price profile of Shangrao city in different directions. The result shows that the residential land price in Shangrao city has a certain agglomeration effect on the space. And Shangrao residential land price is distributed as a "single core" spatial structure, showing a decreasing trend from the center of the city to the periphery, in addition, there are "cold spots" and "hot spots" of land price in the area. Through the selection of weight function and the determination of bandwidth, the GWR model is used to determine the residential land price in Shangrao city. The relationship between residential land price and its influencing factors in Shangrao City is discussed in detail. The results of GWR model are compared and analyzed by variance analysis, event intelligence diagnosis and Monte Carlo significance test. This paper establishes the adjusted spatial core as the final model to explain the spatial structure of residential land price, and analyzes the spatial distribution difference of various influencing factors on residential land price with GIS tool. The results show that the hospital and middle school. The influence factors such as primary school, park, area rate and green rate have significant influence on residential land price, and the influence degree of each influencing factor is different with the change of space position. Finally, the results of the two models are analyzed, and it is found that the GWR model has a better explanation than the feature price model.
【学位授予单位】：江西师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F299.23

【参考文献】