TokenComparator-class

Represents a comparator for pairs of token sequences.

Implements functions for comparing strings, sequences and
numeric vectors for clustering and record linkage applications.
Supported comparison functions include: generalized edit distances
for comparing sequences/strings, Monge-Elkan similarity for fuzzy
comparison of token sets, and L-p distances for comparing numeric
vectors. Where possible, comparison functions are implemented in
C/C++ to ensure good performance.

Neil Marchant

comparator

Comparison Functions for Clustering and Record Linkage

TokenComparator-class function

<dl>
<dt><code>.Data</code></dt>
<dd>a function that calls the elementwise method for this class,
with arguments <code>x</code>, <code>y</code> and <code>...</code>.</dd></dl>
<dt><code>symmetric</code></dt>
<dd>a logical of length 1. If TRUE, the comparator is symmetric
in its arguments---i.e. <code>comparator(x, y)</code> is identical to
<code>comparator(y, x)</code>.</dd>
<dt><code>distance</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator produces
distances and satisfies <code>comparator(x, x) = 0</code>. The comparator may not
satisfy all of the properties of a distance metric.</dd>
<dt><code>similarity</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator produces
similarity scores.</dd>
<dt><code>tri_inequal</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator satisfies
the triangle inequality. This is only possible (but not guaranteed) if
<code>distance = TRUE</code> and <code>symmetric = TRUE</code>.</dd>
<dt><code>ordered</code></dt>
<dd>a logical of length 1. If TRUE, the comparator treats token
sequences as ordered, otherwise they are treated as unordered.</dd>

Slots

Virtual Token Comparator Class — TokenComparator-class

<dl>
<dt><code>.Data</code></dt>
<dd>a function that calls the elementwise method for this class,
with arguments <code>x</code>, <code>y</code> and <code>...</code>.</dd>


<dt><code>symmetric</code></dt>
<dd>a logical of length 1. If TRUE, the comparator is symmetric
in its arguments---i.e. <code>comparator(x, y)</code> is identical to
<code>comparator(y, x)</code>.</dd>


<dt><code>distance</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator produces
distances and satisfies <code>comparator(x, x) = 0</code>. The comparator may not
satisfy all of the properties of a distance metric.</dd>


<dt><code>similarity</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator produces
similarity scores.</dd>


<dt><code>tri_inequal</code></dt>
<dd>a logical of length 1. If <code>TRUE</code>, the comparator satisfies
the triangle inequality. This is only possible (but not guaranteed) if
<code>distance = TRUE</code> and <code>symmetric = TRUE</code>.</dd>


<dt><code>ordered</code></dt>
<dd>a logical of length 1. If TRUE, the comparator treats token
sequences as ordered, otherwise they are treated as unordered.</dd>


</dl>

TokenComparator-class: Virtual Token Comparator Class

Description

Arguments

Slots