Local influence when fitting Gaussian spatial linear models: an agriculture application

D.M. Grzegozewski, M.A. Uribe-Opazo, F. De Bastiani, and M. Galea. 2013. Local influence when fitting Gaussian spatial linear models: an agriculture application. Cien. Inv. Agr. 40(3): 523-535. Outliers can adversely affect how data fit into a model. Obviously, an analysis of dependent data is different from that of independent data. In the latter, i.e., in cases involving spatial data, local outliers can differ from the data in the neighborhood. In this article, we used the local influence technique to identify influential points in the response variables using two different schemes of perturbations. We applied this technique to soil chemical properties and soybean yield. We evaluated the effects of the influential points on the spatial model selection, the parameter estimation by maximum likelihood and the construction of thematic maps by kriging. In the construction of the thematic maps in studies with and without the influential points, there were changes in the levels of nutrients, allowing for the appropriate application of input, generating greater savings for the producer and contributing to the protection of the environment.

[1]  E. Silva,et al.  Seleção de modelos de variabilidade espacial para elaboração de mapas temáticos de atributos físicos do solo e produtividade da soja , 2008 .

[2]  R. Cook.  Assessment of Local Influence , 1986 .

[3]  R. Lark,et al.  Geostatistics for Environmental Scientists , 2001 .

[4]  Local in∞uence of explanatory variables in Gaussian spatial linear models , 2011 .

[5]  A. McBratney,et al.  Choosing functions for semi‐variograms of soil properties and fitting them to sampling estimates , 1986 .

[7]  A. Konopka,et al.  FIELD-SCALE VARIABILITY OF SOIL PROPERTIES IN CENTRAL IOWA SOILS , 1994 .

[8]  K. Mardia,et al.  Maximum likelihood estimation of models for residual covariance in spatial regression , 1984 .

[9]  Dale L. Zimmerman,et al.  A comparison of spatial semivariogram estimators and corresponding ordinary Kriging predictors , 1991 .

[10]  Miguel Angel Uribe-Opazo,et al.  Comparação de mapas de variabilidade espacial da resistência do solo à penetração construídos com e sem covariáveis usando um modelo espacial linear , 2012 .

[11]  R. Lark.  Optimized spatial sampling of soil for estimation of the variogram by maximum likelihood , 2002 .

[12]  R. M. Lark,et al.  Estimating variograms of soil properties by the method‐of‐moments and maximum likelihood , 2000 .

[13]  M. Genton.  Spatial Breakdown Point of Variogram Estimators , 1998 .

[14]  J. Ibrahim,et al.  Perturbation selection and influence measures in local influence analysis , 2007, 0803.2986.

[15]  Liviu Theodor Ene,et al.  Modelling tree diameter from airborne laser scanning derived variables: A comparison of spatial statistical models , 2010 .

[16]  L. H. C. Anjos,et al.  Sistema Brasileiro de Classificação de Solos. , 2006 .

[17]  Miguel Angel Uribe-Opazo,et al.  Influence diagnostics in Gaussian spatial linear models , 2012 .

[18]  R. Reese.  Geostatistics for Environmental Scientists , 2001 .

[19]  A. Hossain,et al.  A comparative study on detection of influential observations in linear regression , 1991 .

[20]  R. Cook.  Detection of influential observation in linear regression , 2000 .

[21]  N. Cressie,et al.  Robust estimation of the variogram: I , 1980 .

[22]  Ronald Christensen,et al.  Covariance function diagnostics for spatial linear models , 1993 .

[23]  R. M. Lark,et al.  A comparison of some robust estimators of the variogram for use in soil survey , 2000 .