Analysis of Large Geostatistical Data Sets
April 6, 2009
4:00 p.m.
Dr. Hao Zhang
Abstract
Geostatistical data refer to those observations that are observed at different spatial locations so that they have a location attribute. Not only the observations but also the spatial attributes are of interest.
Geostatistical data arise in mining and petroleum industries, agricultural and environmental sciences, and many more disciplines. With the advance of technology, the collection of geostatistical data has become easier and in some cases automated. Consequently, a large set of geostatistical data is usually available in a study (as in the study on climate change). One challenge to the statistical inference is the inversion of a large covariance matrix, which is needed in the likelihood-based inference or Bayesian analysis, and in the linear interpolation. In this talk, I will cover some methods to deal with this challenge.