Analysing positive-valued spatial data: the transformed Gaussian model

The Gaussian assumption is often inappropriate for analysing some kinds of geostatistical data and transformations can be used in an attempt to get nearly-Gaussian behaviour. Here we study the transformed Gaussian model, which includes an additional parameter corresponding to the Box-Cox family of transformations. In particular we consider maximum likelihood estimation and minimum mean square error prediction for this model, and as an example we use it to model rainfall data. We discuss the limitations of the transformed Gaussian model, and suggest that it should be used primarily as a first line of attack in dealing with non-Gaussianity and non-linearity, before proceding to more complex models.

This talk is based on joint work with Peter Diggle and Paulo Ribeiro Jr.


Last modified: Thu Nov 2 11:22:16 MET 2000