loglik_HR

Computes (censored) Huesler-Reiss log-likelihood, AIC, and BIC values.

Statistical methodology for sparse multivariate extreme value models. Methods are
provided for exact simulation and statistical inference for multivariate Pareto distributions
on graphical structures as described in the paper 'Graphical Models for Extremes' by
Engelke and Hitz (2020) <doi:10.1111/rssb.12355>.

Sebastian Engelke

graphicalExtremes

Statistical Methodology for Graphical Extreme Value Models

Adrien S. Hitz

Nicola Gnecco

Manuel Hentschel

loglik_HR function

<dl><dt>data</dt>
<dd>Numeric \(n \times d\) matrix. It contains
observations following a multivariate HR Pareto distribution.</dd>
<dt>p</dt>
<dd>Numeric between 0 and 1 or <code>NULL</code>. If <code>NULL</code> (default),
it is assumed that the <code>data</code> are already on multivariate Pareto scale.
Else, <code>p</code> is used as the probability in the function <code>data2mpareto()</code>
to standardize the <code>data</code>.</dd>
<dt>graph</dt>
<dd>An [<code><a href="/link/igraph%3A%3Agraph?package=graphicalExtremes&version=0.3.2" data-mini-rdoc="graphicalExtremes::igraph::graph">igraph::graph</a></code>] object or <code>NULL</code>. The <code>graph</code> must be undirected and
connected. If no graph is specified, the complete graph is used.</dd>
<dt>Gamma</dt>
<dd>Numeric \(n \times d\) matrix.
It represents a variogram matrix \(\Gamma\).</dd>
<dt>cens</dt>
<dd>Boolean. If true, then censored log-likelihood is computed.
By default, <code>cens = FALSE</code>.</dd></dl>

Arguments

Compute Huesler-Reiss log-likelihood, AIC, and BIC — loglik_HR

<dl>

<dt>data</dt>
<dd>Numeric \(n \times d\) matrix. It contains
observations following a multivariate HR Pareto distribution.</dd>


<dt>p</dt>
<dd>Numeric between 0 and 1 or <code>NULL</code>. If <code>NULL</code> (default),
it is assumed that the <code>data</code> are already on multivariate Pareto scale.
Else, <code>p</code> is used as the probability in the function <code>data2mpareto()</code>
to standardize the <code>data</code>.</dd>


<dt>graph</dt>
<dd>An [<code><a href='https://rdrr.io/pkg/igraph/man/make_graph.html'>igraph::graph</a></code>] object or <code>NULL</code>. The <code>graph</code> must be undirected and
connected. If no graph is specified, the complete graph is used.</dd>


<dt>Gamma</dt>
<dd>Numeric \(n \times d\) matrix.
It represents a variogram matrix \(\Gamma\).</dd>


<dt>cens</dt>
<dd>Boolean. If true, then censored log-likelihood is computed.
By default, <code>cens = FALSE</code>.</dd>

</dl>

loglik_HR: Compute Huesler-Reiss log-likelihood, AIC, and BIC

Description

Usage

Value

Arguments

See Also