The function transforms observations belonging to the GEV class from one model to another.
Dist2Dist(data, from='Gev', to='sFrechet', loc=NULL, scale=NULL,
shape=NULL)
A numeric vector or a matrix of extreme values.
The name of the original extreme value distribution,
i.e. Gev
(the default), see the Details section.
The name of the desired extreme value distribution,
i.e. sFrechet
(the default),
see the Details section.
A numeric value or vector of location parameters.
A numeric value or vector of scale parameters.
A numeric value or vector of shape parameters.
A numeric vector or matrix of transformed values following the desired distribution.
If data
is a numeric vector of length n
then the dataset is consider as a
realisation from an univariate extreme value distribution. Instead, if
data
is a (\(n \times d\))-matrix then the columns
represent the different variables with extreme value distributions
and the rows represent the iid replications. Finally,
if data
is a (\(d \times d \times n\))-matrix then
the columns and rows represent the different variables and the third
dimension represents the iid replications.
The parameters from
and to
indicate the original extreme
value distribution(s) from which the observations are drawn and the
target extreme value distribution(s) that the transformed data will
follow. The options are:
from=Gev
(generalised extreme value distribution):
to=Uniform
, which means uniform distribution;
to=sFrechet
, which means standard (or unit) Frechet distribution,
that is GEV(1,1,1);
to=sGumbel
, which means standard Gumbel distribution, that is GEV(0,1,1);
to=sWeibull
, which means standard Weibull distribution, that is GEV(1,1,-1);
to=Gev
, which means generalised extreme value
distribution. Note, that in this case, it is required to insert vectors of
location, scale and shape parameters with dimension n
in the
univariate case, dimension d
when data
is (\(n \times d\))-matrix and dimension \(n \times d\) when data
is
(\(d \times d \times n\))-matrix.
from=sFrechet
to=Gev
.
from=sGumbel
to=Gev
.
from=sWeibull
to=Gev
.
de Haan, L. and Ferreira, A. (2006) Extreme Value Theory An Introduction. Springer Verlang, New York.