Inner View of Spatial Autoregression Residuals with LADS and Join-Count Statistics

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Spatial autocorrelation is a reality but also a requirement to carry out spatial interpolation and spatial autoregressive modeling. Conventional statistics must, then, be reformulated to properly account for spatial correlation and spatial heterogeneity within georeferenced data. For instance, if autoregressive residuals reveal a medium-strong spatial autocorrelation then any omitted variable, within the initial regression model, can be significant. Hence, LADS and join-count statistics are reviewed here as a way to test this missing explanatory variables through its spatial binary residuals because spatial autocorrelation error can be defined as a specification error artifact for spatial modeling.
Original languageUnknown
Title of host publicationConference of the International Statistic Institute
Pages1-5
Publication statusPublished - 1 Jan 2009
Event57th Conference of the International Statistic Institute -
Duration: 1 Jan 2009 → …

Conference

Conference57th Conference of the International Statistic Institute
Period1/01/09 → …

Cite this