Data-Driven Spatial Basis Functions in Fixed Rank Kriging for Large Spatial Data Sets

• Activity Number: 314
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Emily L. Kang*
• Companies: University of Cincinnati
• Keywords: Basis functions ; Fixed Rank Kriging ; Spatial prediction

Abstract

Fixed Rank Kriging (FRK) (Cressie and Johannesson, 2008) has become an effective and computational efficient method for analyzing large spatial data sets. Its implementation usually requires pre-specified spatial basis functions. In this study, we investigate the role of these spatial basis functions in the performance of FRK and propose a data-driven approach of constructing such basis functions. Simulation experiments and real-data analyses are carried out to compare different basis functions and to demonstrate the performance of the suggest data-drive basis functions.