u with
parameters scale sigmau and shape xi.
Unconditional likelihood also provided when the
probability phiu of being above the threshold
u is given.lgpd(x, u = 0, sigmau = 1, xi = 0, phiu = 1, log = TRUE)
nlgpd(pvector, x, u = 0, phiu = 1, finitelik = FALSE)sigmau, xi) or NULLu as used in the maximum likelihood
fitting function fgpd.
They are designed to be used for MLE in
fgpd but are available for
wider usage, e.g. constructing your own extreme value
mixture models.
See fgpd and
dgpd for full details.
Log-likelihood calculations are carried out in
lgpd, which takes parameters as
inputs in the same form as distribution functions. The
negative log-likelihood is a wrapper for
lgpd, designed towards making
it useable for optimisation (e.g. parameters are given a
vector as first input).
Unlike the MLE function fgpd,
the phiu must be in range $[0, 1]$ and cannot
be NULL. Specify phiu=1 for conditional
likelihood (default) and phiu<1< code=""> for unconditional
likelihood.
The function lgpd carries out
the calculations for the log-likelihood directly, which
can be exponentiated to give actual likelihood using
(log=FALSE).evd package.
dgpd, fpot
and fitdistr
Other gpd: dgpd, fgpd,
gpd, pgpd,
qgpd, rgpd