摘要
将空间统计分析应用于空间关联挖掘领域,给出空间权重矩阵、空间自相关和空间关联的度量函数,并以中国有代表性的37个大中城市的地理空间数据为例,进行空间关联研究。根据空间数据的地理位置构造其Voronoi图、Delaunay图,计算空间对象之间的距离并构造其邻域图和空间自相关矩阵,在此基础上计算空间对象间的空间自相关系数和局部空间关联系数,包括Moran’s I、Gereay’s C、局部Moran、G统计,并依据这些系数发现空间对象间的空间关联知识。
Spatial statistical analysis techniques are applied in spatial association mining. The measurement functions of spatial weight matrix, spatial autocorrelation and spatial association are studied. Meanwhile, the experiences concerned are performed using the geographical spatial data gotten from 37 typified cites in China. Voronoi and Delaunay diagram, neighbor graph and spatial autocorrelation matris are founded according to the geographical position and the distance among the spatial objects. On the basis of these methods, spatial autocorrelation coefficients and local spatial association coefficients are computed, including Moran's I, Greary's C, Local Moran, G Statistics and spatial association knowledge is acquired according to these coefficients.
出处
《计算机工程》
EI
CAS
CSCD
北大核心
2006年第5期20-22,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60275021)
关键词
空间统计学
空间自相关
空间权重矩阵
空间关联
Spatial statistics
Spatial autocorrelation
Spatial weight matrix
Spatial association