compute_features

This function computes features for each network using graph theoretic constructs.

Anomaly detection in dynamic, temporal networks. The package
'oddnet' uses a feature-based method to identify anomalies. First, it computes
many features for each network. Then it models the features using time series
methods. Using time series residuals it detects anomalies. This way, the
temporal dependencies are accounted for when identifying anomalies
(Kandanaarachchi, Hyndman 2022) <arXiv:2210.07407>.

Sevvandi Kandanaarachchi

oddnet

Anomaly Detection in Temporal Networks

Rob Hyndman

compute_features function

<dl><dt>gr</dt>
<dd>The network or graph as an <code>igraph</code> object.</dd>
<dt>attributes</dt>
<dd>If the network nodes/vertices have attributes, then <code>attributes = TRUE</code>.</dd>
<dt>attr_name</dt>
<dd>The name of the node/vertex attribute. Only a single attribute can
be specified.</dd>
<dt>fast</dt>
<dd>If set to <code>TRUE</code> will avoid computing time consuming features.</dd></dl>

Arguments

Computes features for each network. — compute_features

<dl>

<dt>gr</dt>
<dd>The network or graph as an <code>igraph</code> object.</dd>


<dt>attributes</dt>
<dd>If the network nodes/vertices have attributes, then <code>attributes = TRUE</code>.</dd>


<dt>attr_name</dt>
<dd>The name of the node/vertex attribute. Only a single attribute can
be specified.</dd>


<dt>fast</dt>
<dd>If set to <code>TRUE</code> will avoid computing time consuming features.</dd>

</dl>

compute_features: Computes features for each network.

Description

Usage

Value

Arguments

Examples