uni_excess: Univariate threshold excess model

Description Usage Arguments

View source: R/uni_excess.R

Description

Fit a standard univariate threshold exceedance model.

Usage

1
2
uni_excess(x, uq = 0.95, threshold, r = 0, m.ksi = 0, s.ksi = 10,
  nburn = 10000, nmcmc = 10000, window = 200, chain_init)

Arguments

x

Numeric vector of data.

uq

numeric, within (0, 1), the quantile of the complete data (includes replicates if available) that determines the threshold

threshold

numeric, values greater than threshold are considered excess. Defaults to the threshold given by uq, but if specified, uq is ignored.

r

Numeric, nonnegative integer, specifies the degree of separation between clusters in a declustering scheme. Defaults to r = 0 (no clustering).

m.ksi, s.ksi

numeric, prior mean and sd for ksi, ksi is normal

nburn

numeric, nonnegative integer, the first nburn samples are discared. Defaults to 80000.

nmcmc

numeric, positive integer, the number of samples that are kept _post_ burn-in. That is, a total of nburn + nmcmc iterations of the chain are run. Defaults to 40000.

window

numeric, during the burn-in, the covariance matrix of the proposal distribution is changed every window iterations to improve the acceptance rate. Also, the iteration number is displayed every window iterations. Defaults to 500.

chain_init

numeric vector of length 2, specifying the starting location of the Markov chain for (sigma, ksi). Defaults to c(1, 1e-6).


mickwar/mwEVT documentation built on May 22, 2019, 9:56 p.m.