excessProb.pb {BMAmevt}

Estimates the probability of joint excess (Frechet margins)

Description

Double Monte-Carlo integration.

Usage

excessProb.pb(
  post.sample,
  Nmin.intern = 100,
  precision = 0.05,
  from = NULL,
  to = NULL,
  thin = 100,
  displ = FALSE,
  thres = rep(500, 5),
  known.par = FALSE,
  true.par
)

Arguments

post.sample

The posterior sample.

Nmin.intern

The minimum number of MC iteration in the internal loop (excess probability, conditional to a parameter).

precision

The desired precision for the internal MC estimate

from

Integer or NULL. If NULL, the default value is used. Otherwise, should be greater than post.sample$Nbin. Indicates the index where the averaging process should start. Default to post.sample$Nbin +1

to

Integer or NULL. If NULL, the default value is used. Otherwise, must be lower than Nsim+1. Indicates where the averaging process should stop. Default to post.sample$Nsim.

thin

Thinning interval.

displ

logical. Should a plot be produced ?

thres

A multivariate threshold

known.par

Logical

true.par

The true parameter from which the data are issued.

Value

A list made of

whole

A vector of estimated excess probabilities, one for each element of the thinned posterior sample.

mean

the estimated threshold excess probability: mean estimate.

esterr

The estimated standard deviation of the mean estimate (where the Monte-Carlo error is neglected)

estsd

The estimated standard deviation of the posterior sample (where the Monte-Carlo error is neglected)

lowquants

The three lower 0.1 quantiles of, respectively, the conditional mean estimates and of the upper and lower bounds of the Gaussian (centered) 80 % confidence intervals around the conditional estimates.

upquants

The three upper 0.9 quantiles

true.est

the mean estimate conditional to the true parameter: a vector of size three: the mean estimate , and the latter +/- the standard deviation of the estimate