摘要
研究了最常用的衡量全局空间自相关的指标:全局Moran指数和全局G系数,基于模拟的空间区域,设计了一些有代表性的空间聚集方案进行计算,比较了两种指标的探测结果。得出的初步结论是:G系数探测高值聚集的能力强于低值聚集;当研究范围内同时存在高值和低值聚集时,G系数会受聚集区域规模的影响,当高值聚集区域和低值聚集区域规模相当时,G系数往往为正数,表明G系数对高值敏感;Moran指数主要受聚集区域规模的影响,随着空间聚集范围的扩展,Moran指数会明显增大。这些结论对于探测空间自相关模式时选择何种全局指标有指导意义。
Moran＇s Index and Getis-Ord general G are among the most commonly used indices for the analysis of general spatial autocorrelation.This paper compared Moran＇s Index and Getis-Ord General G,based on simulating computation,the results showed that：Getis-Ord general G is more sensitive to high clusters than low ones.When high and low cluster coexist at the study area,Getis-Ord general G will be affected by the scale of cluster.G will more likely to be positive to indicate high cluster when the scale of high and low cluster are equal;Moran＇s Index is mainly affected by the scale of clusters,it will increase when the scale is extended.These conclusions are useful in selecting general index to detect the spatial autocorrelation pattern.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第4期93-97,共5页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
国家自然科学基金资助项目(40604001)
现代工程测量国家测绘局重点实验室开放基金(ESSBSM0611)
关键词
空间自相关
全局指标
MORAN指数
G系数
spatial autocorrelation
general indices
Moran＇s Index
Getis-Ord general G
