Spatial prediction for axially symmetric process on spheres
- UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
- Sarangan Balasubramaniam (Creator)
- Institution
- The University of North Carolina at Greensboro (UNCG )
- Web Site: http://library.uncg.edu/
- Advisor
- Haimeng Zhang
Abstract: Spatial prediction, or so-called kriging, is one of the ultimate goals in spatial data analysis. The basic idea of kriging is to use the values of a geographic variable at some locations to estimate the value(s) that are unknown at other locations. In this dissertation, we consider the spatial prediction when a random process is axially symmetric on the sphere. More specifically, we first decompose an axially symmetric process as Fourier series on circles, where the Fourier random coefficients can be expressed as circularly-symmetric complex random processes. The estimation of covariance functions for complex random processes is then obtained through both parametric and non-parametric approaches, where least squared error estimation and the Wavelet-Galerkin methods are applied, respectively. Ordinary kriging is then conducted on possibly complex random fields and predicted data values are computed through the inverse Discrete Fourier transformation. All the above approaches and results are demonstrated through simulation studies.
Spatial prediction for axially symmetric process on spheres
PDF (Portable Document Format)
1139 KB
Created on 8/1/2022
Views: 169
Additional Information
- Publication
- Dissertation
- Language: English
- Date: 2022
- Keywords
- Kriging, Spatial data analysis, Axially symmetric process
- Subjects
- Spatial analysis (Statistics)
- Kriging