Regularization Methods for the Selection of Covariates and Dependence Structure in Spatial Models

• Activity Number: 379
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Jun Zhu*+
• Companies: University of Wisconsin
• Keywords: autoregression ; covariance function ; regression analysis ; spatial statistics ; variable selection

Abstract

Regularization methods can be applied for variable selection in spatial models. The complexity of computation, methods, and theory may differ greatly depending on the objectives of the studies. Here I contrast methods for the selection of covariates, which tends to be more compatible with those for non-spatial models, against those for the selection of spatial dependence structure. While the primary focus is on Gaussian spatial linear regression models, non-Gaussian spatial models are also considered. Examples of ecological data are given for illustration where the selection of covariate and spatial dependence structure are of interest.