Damijana Kastelec wrote:
>
> Hi!
> I am doing cokriging on precipitation data, I would like to do two
> transformation on the data: log and standard, but I don't know what
> happend so:
> - Is it possible to use both transformation?
I think it is (though I cannot think of any use for this)
> - Are the data first transformed with log transformation and than
> standardized?
Yes. Note that the standardization only implies division by the
(data) standard deviation.
> - Which value is used as mean and standard deviation for
> standardization?
So,.. no subtraction of the mean.
> - What happened by cross validation, what are the predicted
> values?
Predictions at the data locations, leaving the value at the prediction
location itself out. It's only done for the first variable in the
command file.
> - In cross validation I only get predictions and prediction variances
> for the first variable, how I can get this also for the second variable
> and the cros-covariances?
>
Switch the order of the `data(var): ... ;' statements your command
file starts with. Cross-covariances are not cross validated themselves,
however they are taken into account during cross validation (by using
cokriging for the prediction)
--
Edzer
