Random sampling from D-gaps posterior distribution

Uses the rust package to simulate from the posterior distribution of the extremal index $\theta$ based on the D-gaps model for threshold interexceedance times of Holesovsky and Fusek (2020). We refer to this as the $D$-gaps model, because it uses a tuning parameter $D$, whereas the related $K$-gaps model of Suveges and Davison (2010) has a tuning parameter $K$.

Usage

dgaps_post(
  data,
  thresh,
  D = 1,
  n = 1000,
  inc_cens = TRUE,
  alpha = 1,
  beta = 1,
  param = c("logit", "theta"),
  use_rcpp = TRUE
)

Arguments

data

A numeric vector or numeric matrix of raw data. If data is a matrix then the log-likelihood is constructed as the sum of (independent) contributions from different columns. A common situation is where each column relates to a different year.

If data contains missing values then split_by_NAs is used to divide the data further into sequences of non-missing values, stored in different columns in a matrix. Again, the log-likelihood is constructed as a sum of contributions from different columns.

thresh

A numeric scalar. Extreme value threshold applied to data.

D

A numeric scalar. The censoring parameter $D$, as defined in Holesovsky and Fusek (2020). Threshold inter-exceedances times that are not larger than D units are left-censored, occurring with probability $\log(1 - \theta e^{-\theta d})$, where $d = q D$ and $q$ is the probability with which the threshold $u$ is exceeded.

n

A numeric scalar. The size of posterior sample required.

inc_cens

A logical scalar indicating whether or not to include contributions from right-censored inter-exceedance times, relating to the first and last observations. It is known that these times are greater than or equal to the time observed. If data has multiple columns then there will be right-censored first and last inter-exceedance times for each column. See also the Details section of dgaps.

alpha, beta

Positive numeric scalars. Parameters of a beta($\alpha$, $\beta$) prior for $\theta$.

param

A character scalar. If param = "logit" (the default) then we simulate from the posterior distribution of $\phi = \log(\theta / (1-\theta))$ and then transform back to the $\theta$-scale. If param = "theta" then we simulate directly from the posterior distribution of $\theta$, unless the sample D-gaps are all equal to zero or all positive, when we revert to param = "logit". This is to avoid the possibility of sampling directly from a posterior with mode equal to 0 or 1.

use_rcpp

A logical scalar. If TRUE (the default) the rust function ru_rcpp is used for posterior simulation. If FALSE the (slower) function ru is used.

Value

An object (list) of class "evpost", which has the same structure as an object of class "ru" returned from ru. In addition this list contains

call: The call to dgaps().
model: The character scalar "dgaps".
thresh: The argument thresh.
ss: The sufficient statistics for the D-gaps likelihood, as calculated by dgaps_stat.

Details

A beta($\alpha$, $\beta$) prior distribution is used for $\theta$ so that the posterior from which values are simulated is proportional to $$\theta ^ {2 N_1 + \alpha - 1} (1 - \theta e^{-\theta d}) ^ {N_0 + \beta - 1} \exp\{- \theta q (I_0 T_0 + \cdots + I_N T_N)\}.$$ See dgaps_stat for a description of the variables involved in the contribution of the likelihood to this expression.

The ru function in the rust package simulates from this posterior distribution using the generalised ratio-of-uniforms distribution. To improve the probability of acceptance, and to ensure that the simulation will work even in extreme cases where the posterior density of $\theta$ is unbounded as $\theta$ approaches 0 or 1, we simulate from the posterior distribution of $\phi = \log(\theta / (1-\theta))$ and then transform back to the $\theta$-scale.

References

Holesovsky, J. and Fusek, M. Estimation of the extremal index using censored distributions. Extremes 23, 197-213 (2020). doi:10.1007/s10687-020-00374-3

Suveges, M. and Davison, A. C. (2010) Model misspecification in peaks over threshold analysis, The Annals of Applied Statistics, 4(1), 203-221. doi:10.1214/09-AOAS292

Examples

# Newlyn sea surges

thresh <- quantile(newlyn, probs = 0.90)
d_postsim <- dgaps_post(newlyn, thresh)
plot(d_postsim)


### Cheeseboro wind gusts

d_postsim <- dgaps_post(exdex::cheeseboro, thresh = 45, D = 3)
plot(d_postsim)