simDAG

Simulates a random directed acyclic graph adjacency (DAG) matrix with
the provided edge sparsity. The edge sparsity is the percentage of edges
that are absent, relative to a fully connected DAG.

Various tools for inferring causal models from observational data. The package
includes an implementation of the temporal Peter-Clark (TPC) algorithm. Petersen, Osler
and Ekstrøm (2021) <doi:10.1093/aje/kwab087>. It also includes general tools
for evaluating differences in adjacency matrices, which can be used for evaluating
performance of causal discovery procedures.

Bjarke Hautop Kristensen

causalDisco

Tools for Causal Discovery on Observational Data

Anne Helby Petersen

Tobias Ellegaard Larsen

Claus Thorn Ekstrøm

simDAG function

<dl><dt>p</dt>
<dd>The number of nodes.</dd>
<dt>sparsity</dt>
<dd>If <code>NULL</code> (the default), a random edge sparsity
is sampled from the interval provided in <code>sparsityLim</code>. Otherwise,
the sparsity should be provided as a numeric in [0,1].</dd>
<dt>sparsityLim</dt>
<dd>A vector of two numerics, both must be in [0,1].</dd>
<dt>permute</dt>
<dd>If <code>FALSE</code>, the adjacency matrix will include nodes
in their causal ordering. This is avoided by setting <code>permute = TRUE</code>,
in which case the node order is permuted randomly.</dd></dl>

Arguments

Simulate a random DAG — simDAG

<dl>

<dt>p</dt>
<dd>The number of nodes.</dd>


<dt>sparsity</dt>
<dd>If <code>NULL</code> (the default), a random edge sparsity
is sampled from the interval provided in <code>sparsityLim</code>. Otherwise,
the sparsity should be provided as a numeric in [0,1].</dd>


<dt>sparsityLim</dt>
<dd>A vector of two numerics, both must be in [0,1].</dd>


<dt>permute</dt>
<dd>If <code>FALSE</code>, the adjacency matrix will include nodes
in their causal ordering. This is avoided by setting <code>permute = TRUE</code>,
in which case the node order is permuted randomly.</dd>

</dl>

simDAG: Simulate a random DAG

Description

Usage

Value

Arguments

Examples