[en] Distinguishing the analysis of spatial data from classical analysis is only meaningful if
the spatial components bring information. Therefore, testing if the spatial autocorrelation
is significant may confirm or deny the need to consider spatial analysis over the classical
one. Spatial autocorrelation expresses the dependence between values at neighbouring
locations. Several measures of spatial autocorrelation are defined in the literature. Moran’s
index, Geary’s ratio and Getis-Ord statistic are the most used statistics. Tests based on
these measures have been developed in the literature using asymptotic and permutation
results. They are used in practice in many fields, for instance in geography, economics,
biogeosciences, medicine, ... However, these tests should be cautiously applied because they
are not robust. A single contaminated observation can significantly modify their results.
The talk has two main objectives. Firstly, the already available tools for measuring spatial
autocorrelation will be reviewed with an emphasis on the study and comparison of their
robustness. Secondly, alternative methods will be proposed to robustly estimate the spatial
autocorrelation.
Disciplines :
Mathematics
Author, co-author :
Ernst, Marie ; Université de Liège > Département de mathématique > Statistique mathématique
Haesbroeck, Gentiane ; Université de Liège > Département de mathématique > Statistique mathématique
Language :
English
Title :
Spatial autocorrelation: robustness of measures and tests
Publication date :
14 December 2015
Event name :
8th International Conference of the ERCIM WG on Computational and Methodological Statistics (CMStatistics 2015)
Event organizer :
ERCIM Working Group on Computational and Methodological Statistics (CMStatistics), Queen Mary University of London, Birkbeck University of London and Imperial College London.
