In proto4426/PissoortThesis: Thesis in Extreme Value Theory

knitr::opts_chunk$set(echo = TRUE)

R code

• The application is hosted at the following URL .
• The global repository for the thesis which contains the R package can be found here
• The specific code for this app can be found in this folder.

The following package are needed to run the application :

install.packages(c("shiny", "htmltools", "shinythemes","tidyverse","DT",
                   "gridExtra","grid","mvtnorm","plotly","HDInterval","ggjoy","scales","heatmaply", "coda", "reshape2", "ggcorrplot", "viridis"))
## And of course the package for this thesis ... 
devtools::install_github("proto4426/PissoortThesis", build_vignettes=T)

Model

• The number of starting values (see diagnostics) influence the total number of iterations !
• The seed is set for both the random generation of the starting values (see Algorithm 1 in the text) and the random generation of all the Markov chains. All results can thus be easily retrieved, without having to set two different seeds.

Prior

• Yet, these priors are taken to be univariate Normal distributions for each parameters. It would be interesting to incorporate other forms of priors as discussed in
• The values set by default here and in the thesis are chosen... by default ! But, since these are taken to be uninformative (near-flat) priors, relatively small changes won't change the ouput. Verify it by your mouse !

Diagnostics

The diagnostics presented in Section 4.4 and used in Section 7.4.1 can be used to assess the chains' convergence to their target stationary distribution.

Posterior Predictive

Presented in Section 4.6 and used in Section 7.4.3, the Posterior Predictive Distribution can be conveniently visualized here for the chosen model.

proto4426/PissoortThesis documentation built on May 26, 2019, 10:31 a.m.