Description

This satellite workshop will review R implementations of a variety of techniques for statistical analysis of extreme values. The focus of the first part of the workshop will be on univariate extremes, including likelihood-based, Bayesian and nonparametric methods for both peaks-over-threshold and block maxima approaches, with a foray into nonstationary extremes. The second part of the workshop will concentrate on conditional extremes and time series.

The tutorial takes place Friday, June 30th, 2023, from 14:00 until 18:15 in room Aula Info AS04.

Course content

Exercices

Instructions

We will be using multiple R packages from the Comprehensive R Archive Network, as well as development versions of some packages.

If you plan on using your own laptop, download and install R (current version 4.3.0, nicknamed “Already Tomorrow”) and an integrated development environment such as RStudio.

devpkgs <- c("lbelzile/mev", "lbelzile/rbm")
cranpkgs <- c("lite", "revdbayes", "texmex", "POT", 
              "coda", "gridExtra", "patchwork",
              "GGally", "extRemes", "evd", "extremogram", 
              "threshr", "tea", "evt0", "remotes",
              "xts", "lubridate", "ggplot2", "exdex")
install.packages(cranpkgs)
remotes::install_github(devpkgs)

You may need to download the R tool chain, the clang and gfortran on Mac to build packages.

If you cannot download the package mev, use the CRAN version via install.packages("mev") and download the data from here.

Instructors