Re: [ai-geostats] normality of data
> Hello list,Sure:
> I know that normality is not required to do geostatistical analysis of
> (georeferenced) data. Still I'm unable to find any publication that
> refers to this statement to which i can refer.
> Any suggestions?
Ecker MD. 2003. Geostatistics: past, present and future. In
Environmetrics, edited by El-Shaarawi AH and Jureckova J, In Encyclopedia
of Life Support Systems (EOLSS), Developed under the auspices of UNESCO,
EOLSS Publishers, Oxford, UK (http://www.eolss.net).
Maybe you can still find a pre-print pdf version of Ecker's article.
Otherwise you would have to subscribe to the EOLSS encyclopedia or write
an email to the author.
In the broader statistical context, geostatistics as explained in most (or
all) textbooks and as practised and explained in this list, belongs to the
branch of distribution-free statistics. Note though that this need not be
so. There is a more, say mainstream, geostatistics that depends on a
formal probability model. As far as I know, as of now it covers the
Gaussian model, as well as the binomial (for binary data) and Poisson (for