qqgpd {tea} | R Documentation |
QQ-Plot against the generalized Pareto distribution for given number of exceedances
Description
Plots the empirical observations above a given threshold against the theoretical quantiles of a generalized Pareto distribution.
Usage
qqgpd(data, nextremes, scale, shape)
Arguments
data |
vector of sample data |
nextremes |
number of exceedances |
scale |
scale parameter of GPD |
shape |
shape parameter of GPD |
Details
If the fitted GPD model provides a reasonable approximation of the underlying sample data the empirical and theoretical quantiles should coincide. So plotting them against each other should result in a straight line. Deviations from that line speak for a bad model fit and against a GPD assumption.
Value
The straight red line gives the line of agreement. The dashed lines are simulated 95 percent confidence intervals. Therefor the fitted GPD model is simulated 1000 times using Monte Carlo. The sample size of each simulation equals the number of exceedances.
Examples
data=rexp(1000) #GPD with scale=1, shape=0
qqgpd(data,1000,1,0)