simmcpot: Simulate an Markov Chain with a Fixed Extreme Value...

In POT: Generalized Pareto Distribution and Peaks Over Threshold

Simulate an Markov Chain with a Fixed Extreme Value Dependence from a Fitted mcpot Object

Description

Simulate a synthetic Markov chain from a fitted 'mcpot' object.

Usage

simmcpot(object, plot = TRUE, ...)

Arguments

object	An object of class 'mcpot'; most often the returned object of the fitmcgpd function.
plot	Logical. If TRUE (the default), the simulated Markov chain is plotted.
...	Other optional arguments to be passed to the plot function.

Details

The simulated Markov chain is computed as follows:

1. Simulate a Markov chain prob with uniform margins on (0,1) and with the fixed extreme value dependence given by object;

2. For all prob such as prob \leq 1 - pat, set mc = NA (where pat is given by object$pat);

3. For all prob such as prob \geq 1 - pat, set prob2 = \frac{prob - 1 + pat}{pat}. Thus, prob2 are uniformly distributed on (0,1);

4. For all prob2, set mc = qgpd(prob2, thresh, scale, shape), where thresh, scale, shape are given by the object$threshold, object$param["scale"] and object$param["shape"] respectively.

Value

A Markov chain which has the same features as the fitted object. If plot = TRUE, the Markov chain is plotted.

Author(s)

Mathieu Ribatet

See Also

fitmcgpd, simmc

Examples

data(ardieres)
flows <- ardieres[,"obs"]

Mclog <- fitmcgpd(flows, 5)
par(mfrow = c(1,2))
idx <- which(flows <= 5)
flows[idx] <- NA
plot(flows, main = "Ardieres Data")
flowsSynth <- simmcpot(Mclog, main = "Simulated Data")

POT documentation built on Oct. 17, 2024, 5:06 p.m.