compare

Comparison of two sets of clusters

Contains functions to simplify the use of data mining methods (classification, regression, clustering, etc.), for students and beginners in R programming. Various R packages are used and wrappers are built around the main functions, to standardize the use of data mining methods (input/output): it brings a certain loss of flexibility, but also a gain of simplicity. The package name came from the French "Fouille de Données en Master 2 Informatique Décisionnelle".

Alexandre Blansch<c3><a9>

fdm2id

Data Mining and R Programming for Beginners

Alexandre Blansché

compare function

<dl><dt>clus</dt>
<dd>The extracted clusters.</dd>
<dt>gt</dt>
<dd>The real clusters.</dd>
<dt>eval</dt>
<dd>The evaluation criterion.</dd>
<dt>comp</dt>
<dd>How the two partitions are compared: <code>"max"</code> (each cluster matched with the
class it agrees with most, averaged over clusters), <code>"cluster"</code> (the same scores, not
averaged) or <code>"pairwise"</code> (every pair of observations, labels ignored). In
<code>"pairwise"</code> mode the three criteria are three classical indices built on the same pair
counts: <code>"accuracy"</code> is the Rand index, <code>"jaccard"</code> the Jaccard index on pairs,
and <code>"kappa"</code> Cohen's kappa on the fourfold table of pair agreements (which, by
Warrens (2008), is also the adjusted Rand index). See <code>compare.accuracy</code>,
<code>compare.jaccard</code> and <code>compare.kappa</code>.</dd></dl>

Arguments

Comparison of two sets of clusters — compare

<dl>

<dt>clus</dt>
<dd>The extracted clusters.</dd>


<dt>gt</dt>
<dd>The real clusters.</dd>


<dt>eval</dt>
<dd>The evaluation criterion.</dd>


<dt>comp</dt>
<dd>How the two partitions are compared: <code>"max"</code> (each cluster matched with the
class it agrees with most, averaged over clusters), <code>"cluster"</code> (the same scores, not
averaged) or <code>"pairwise"</code> (every pair of observations, labels ignored). In
<code>"pairwise"</code> mode the three criteria are three classical indices built on the same pair
counts: <code>"accuracy"</code> is the Rand index, <code>"jaccard"</code> the Jaccard index on pairs,
and <code>"kappa"</code> Cohen's kappa on the fourfold table of pair agreements (which, by
Warrens (2008), is also the adjusted Rand index). See <code>compare.accuracy</code>,
<code>compare.jaccard</code> and <code>compare.kappa</code>.</dd>

</dl>

compare: Comparison of two sets of clusters

Description

Usage

Value

Arguments

See Also

Examples