城市地价空间系统的分形特征提取模型与实证  

Empirical research on fractal feature extraction models of urban land price space system

作　　者：耿槟[1] 朱道林[1] 梁颖[2] 

机构地区：[1]中国农业大学土地资源管理系,北京100193 [2]义乌市国土资源局,义乌310003

出　　处：《系统工程理论与实践》2013年第5期1217-1224,共8页Systems Engineering-Theory & Practice

基　　金：公益性行业科研专项经费(201211001);国家自然科学基金(41001372);中央高校基本科研业务费专项资金(KYCX2011037)

摘　　要：针对城市地价空间系统的特点,运用Moran's I指数,Getis-Ord General G指数和盒维数模拟原理构建地价空间系统分形特征的提取模型.采用北京市414个住宅用地出让数据作为样本进行实证分析.研究发现在3500m范围内,北京市居住地价空间系统存在较强正自相关的分形特征,随着空间尺度增大空间相关性逐步降低;北京市居住地价空间系统呈现低度集聚分布模式,即地价较高样点趋于和地价较高样点相邻,地价较低样点趋于地价较低样点相邻规律;北京市居住地价空间系统分形维数处于2～3之间且均为分数,维数越大地价空间系统越复杂.理论与实证研究表明城市地价空间系统的确存在分形特征,城市地价空间系统分形特征的判定为深入分析地价空间分布,以及进一步的地价空间分形插值提供了重要的理论依据.Considering characteristics of urban land price space system, fractal feature extraction models of urban land price space system based on Moran＇s I index, Getis-Ord General G index and the box counting dimension principle were proposed. The experiment was carried out with residential land market data of Beijing. The results show that there are significant positive autocorrelations for residential land price space system within 3500m. Spatial autocorrelation degree of land price decreases with the increase of spatial scale. Residential land prices are clustered in space such that higher land price tends to be surrounded by higher land price neighbors and lower land price by lower land price neighbors. Further, fractal dimension of residential land price space system is between 2 and 3. The bigger fractal dimension, the more complicated land price space system. The empirical results suggest the space system. It helps us not only for knowing the land exploring fractal interpolation of land price space. existence of fractal features in urban land price price distribution in space better, but also for

关 键 词：城市地价 空间系统 分形特征 提取模型 

分 类 号：F061.6[经济管理—政治经济学] F129.9