Spatial Distribution Based on Semivariogram Model

Gandi Pawitan

Economic Journal of Emerging Markets
Volume 1 Issue 1, 2009

This article aims to discuss some aspects in conducting inferential analysis of census data.In this analysis, the assumptions of normality and IID (independently and identically distribution)in the observations are no longer realistic. Hence conventional analyses which arebased on these assumptions are invalid and unreliable. Other alternatives can be considered,such as semivariogram analysis. Semivariogram analysis assumes that observationsare dependent geographically. The analysis is useful in understanding spatial distribution ofcharacteristics under investigation.Keywords: census, aggregation, semivariogram, autocorrelation, spatial distribution