面向离散点的空间权重矩阵生成算法与实证研究  

Analysis and Implement of Spatial Weight Matrix Based on Discrete Points

作　　者：刘仲刚[1] 李满春[1] 刘剑锋 孙燕[3] 

机构地区：[1]南京大学城市与资源学系,江苏南京210093 [2]河北省地理科学研究所,河北石家庄050011 [3]河海大学移民研究中心,江苏南京210098

出　　处：《地理与地理信息科学》2006年第3期53-56,共4页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(40371091);国土资源部土地资源监测调查工程(2005-6.1-6)

摘　　要：采用阈值法和k-近邻法度量空间上离散点间的空间邻接关系,针对不同的距离计算方式(欧式距离和曼哈顿距离)设计了面向离散点的空间权重矩阵生成算法,使用C#语言在计算机上实现。用该算法对收集的8 367个常州市地价样点构建了不同土地用途地价样点的空间权重矩阵,并计算出分用途的常州市城市地价空间自相关指数。There are many spatial phenomena based on discrete points. Constructing an adjacency spatial weight matrix is the primary process to deal with further spatial analysis and statistics. In this article, the threshold algorithm and k- nearest neighbor algorithm were employed to measure the spatial connectivity among discrete spatial points. For different types of distance, namely Euclidean and Manhattan distance, corresponding algorithms to generate the connectivity weight matrix were advanced, which were implemented with C# programming language.Then,8 367 land price sample points from land price survey in Changzhou City were chosen in the case study. These sample points are first categorized by different sorts of land use, such as residential, industrial and commercial, and the spatial weight matrix is built for each land use type.And also the spatial autocorrelation indices were figured out. Comparison indicates that the algorithms advanced in this artide can measure the spatial point pattern quickly and properly.

关 键 词：空间权重矩阵 空间自相关 离散点 城市地价 常州 

分 类 号：P208[天文地球—地图制图学与地理信息工程] F301.3[天文地球—测绘科学与技术]