Reducing fluctuations in the sample variogram |
| |
Authors: | Xavier Emery |
| |
Institution: | (1) Department of Mining Engineering, University of Chile, Avenida Tupper 2069, Santiago, Chile |
| |
Abstract: | In the analysis of regionalized data, irregular sampling patterns are often responsible for large deviations (fluctuations)
between the theoretical and sample semi-variograms. This article proposes a new semi-variogram estimator that is unbiased
irrespective of the actual multivariate distribution of the data (provided an assumption of stationarity) and has the minimal
variance under a given multivariate distribution model. Such an estimator considerably reduces fluctuations in the sample
semi-variogram when the data are strongly correlated and clustered in space, and proves to be robust to a misspecification
of the multivariate distribution model. The traditional and proposed semi-variogram estimators are compared through an application
to a pollution dataset. |
| |
Keywords: | Spatial statistics Variogram inference Weighted variogram Noncentered covariance Variogram declustering |
本文献已被 SpringerLink 等数据库收录! |
|