gpd.fitrange: Fitting the GPD Model Over a Range of Thresholds

Description Usage Arguments Value See Also Examples


Maximum-likelihood fitting for a stationary GPD model, over a range of thresholds. Graphs of parameter estimates which aid the selection of a threshold are produced.


gpd.fitrange(data, umin, umax, nint = 10, show = FALSE, ...)



A numeric vector of data to be fitted.

umin, umax

The minimum and maximum thresholds at which the model is fitted.


The number of fitted models.


Logical; if TRUE, print details of each fit.


Optional arguments to


Two graphs showing maximum likelihood estimates and confidence intervals of the shape and modified scale parameters over a range of thresholds are produced. A list object is returned invisibly with components: 'threshold' numeric vector of length 'nint' giving the thresholds used, 'mle' an 'nint X 3' matrix giving the maximum likelihood parameter estimates (columns are location, scale and shape respectively), 'se' an 'nint X 3' matrix giving the estimated standard errors for the parameter estimates (columns are location, scale and shape, resp.), 'ci.low', 'ci.up' 'nint X 3' matrices giving the lower and upper 95 intervals, resp. (columns same as for 'mle' and 'se').

See Also, mrl.plot,, pp.fitrange


## Not run: data(rain)
## Not run: gpd.fitrange(rain, 10, 40)

Example output

Loading required package: mgcv
Loading required package: nlme
This is mgcv 1.8-20. For overview type 'help("mgcv-package")'.

ismev documentation built on May 11, 2018, 1:03 a.m.