Performs likelihood-Based inference for stationary time series extremes. The general approach follows Fawcett and Walshaw (2012). Marginal extreme value inferences are adjusted for cluster dependence in the data using the methodology in Chandler and Bate (2007), producing an adjusted log-likelihood for the model parameters. A log-likelihood for the extremal index is produced using the K-gaps model of Suveges and Davison (2010). These log-likelihoods are combined to make inferences about return levels.
The main functions are flite and blite,
which perform frequentist and Bayesian inference for time series extremes,
respectively.
See the vignettes
vignette("lite-1-frequentist", package = "lite") and vignette("lite-2-bayesian", package = "lite")
for an overview of the package.
Chandler, R. E. and Bate, S. (2007). Inference for clustered. data using the independence loglikelihood. Biometrika, 94(1), 167-183. doi:10.1093/biomet/asm015
Fawcett, L. and Walshaw, D. (2012), Estimating return levels from serially dependent extremes. Environmetrics, 23, 272-283. doi:10.1002/env.2133
Suveges, M. and Davison, A. C. (2010) Model misspecification in peaks over threshold analysis, Annals of Applied Statistics, 4(1), 203-221. doi:10.1214/09-AOAS292
flite for frequentist threshold-based inference for
time series extremes.
returnLevel for frequentist threshold-based inference
for return levels.
blite for Bayesian threshold-based inference for
time series extremes.
predict.blite for predictive inference for the
largest value observed in \(N\) years.