【转】Geary's C

李旭，Matlab: Geary's C 原文地址

Introduction

Geary's C is a measure of spatial autocorrelation or an attempt to determine if adjacent observations of the same phenomenon are correlated. Geary's C is also known as Geary's contiguity ratio or simply Geary's ratio. While Moran's I measures global spatial autocorrelation, Geary's C is more sensitive to local spatial autocorrelation. Geary's C is related to Moran's I, but it is not identical.

Geary's C is defined as

where x_i denotes the observed value at location i, xbar is the mean of the variable x over the n locations and w_ij are the elements of the spatial weights matrix, defined as 1 if location i is contiguous to location j and 0 otherwise.

Given the null hypothesis is one of no global spatial autocorrelation, the expected value of Geary's C equals 1. As with Moran's I, inference is based on z-scores:

 

The p-value for the null hypothesis is given as

The value of Geary's C lies between 0 and 2. 1 means no spatial autocorrelation. Values lower than 1 demonstrate increasing positive spatial autocorrelation, whilst values higher than 1 illustrate increasing negative spatial autocorrelation.

Example

以四张NDVI数据为例，代码测试结果如Fig. 1，与SpaceStat返回结果一致。

Fig. 1 

References

[1] Moran's I, fromWikipedia, the free encyclopedia.

[2] Geary's C, from Wikipedia, the free encyclopedia.

[3] Global Spatial Autocorrelation Indices - Moran's I, Geary's C and the General Cross-ProductStatistic. (under the normality assumption the variance, the variance under Randomization方差公式存在两种分布假设：正态分布和随机分布)

[4] BioMedware SpaceStat Help - Geary's C Tutorial.

[5] Geary's C – Springer.

[6] Local Spatial Statistics - ENVI.

[7] Geary's C - AURIN.

[8] Global spatial autocorrelation.

[9] SpaceStat - BioMedware.

[10] Z-score formula for Geary's C, from Applied Data-Centric Social Sciences: Concepts,Data, Computation, and Theory.

 

没有整理与归纳的知识，一文不值！高度概括与梳理的知识，才是自己真正的知识与技能。 永远不要让自己的自由、好奇、充满创造力的想法被现实的框架所束缚，让创造力自由成长吧！ 多花时间，关心他（她）人，正如别人所关心你的。理想的腾飞与实现，没有别人的支持与帮助，是万万不能的。

  本文转自wenglabs博客园博客，原文链接：http://www.cnblogs.com/arxive/p/5087125.html ，如需转载请自行联系原作者