plot.evmOpt: Plots for evmOpt objects

In texmex: Statistical Modelling of Extreme Values

Description

Various plots for evmOpt objects. These differ depending on whether or not there are covariates in the model. If there are no covariates then the diagnostic plots are PP- and QQ-plots, a return level plot (produced by plotrl.evmSim) and a histogram of the data with superimposed density estimate. These are all calculated using the data on the original scale. If there are covariates in the model then the diagnostics consist of PP- and QQ- plots calculated by using the model residuals (which will be standard exponential devaiates under the GPD model and standard Gumbel deviates under the GEV model), and plots of residuals versus fitted model parameters.

Usage

## S3 method for class 'evmOpt'
plot(
  x,
  main = rep(NULL, 4),
  xlab = rep(NULL, 4),
  nsim = 1000,
  alpha = 0.05,
  ...
)

Arguments

x	an object of class evmOpt
main	titles for diagnostic plots. Should be a vector of length 4, with values corresponding to the character strings to appear on the titles of the pp, qq, return level, and density estimate plots respectively.
xlab	As for main but labels for x-axes rather than titles.
nsim	The number of replicates to be simulated to produce the simulated tolerance intervals.
alpha	A 100(1 - alpha)% simulation envelope is produced.
...	FIXME

Details

The PP- and QQ-plots show simulated pointwise tolerance intervals. The region is a 100(1-alpha)% region based on nsim simulated samples.

See Also

evm

texmex documentation built on Dec. 4, 2020, 5:08 p.m.