lowerBound

a <code>matrix</code>, which is either the output of <code><a rd-options="" href="/link/evaluationMatrix?package=RankAggregator&version=0.0.1" data-mini-rdoc="RankAggregator::evaluationMatrix">evaluationMatrix</a></code>,
or a subset of the output of <code><a rd-options="" href="/link/evaluationMatrix?package=RankAggregator&version=0.0.1" data-mini-rdoc="RankAggregator::evaluationMatrix">evaluationMatrix</a></code>.

umat

a vector of rank positions
for each <code>Item</code> in <code>umat</code> that is ranked so far;
partial rankings may contain some - or all - <code>NA</code>
values.

partial.ranking

This function is called by <code><a rd-options="" href="/link/RankAggregator?package=RankAggregator&version=0.0.1" data-mini-rdoc="RankAggregator::RankAggregator">RankAggregator</a></code>::<code><a rd-options="" href="/link/consensusRanking?package=RankAggregator&version=0.0.1" data-mini-rdoc="RankAggregator::consensusRanking">consensusRanking</a></code>.
The lower bound is the absolute lowest value a complete candidate ranking
could attain. Note, this value is not always achievable, so may differ
from the value returned by <code><a rd-options="" href="/link/upperBound?package=RankAggregator&version=0.0.1" data-mini-rdoc="RankAggregator::upperBound">upperBound</a></code>.<pre> For each pair of \code{Item}s, there are three possible calculations, depending
 on whether both \code{Item}s are in the \code{partial.ranking}, one is in
 and the other is out the \code{partial.ranking}, or both are not in
 the \code{partial.ranking}.
</pre>

Easily compute an aggregate ranking (also called a median ranking or a
consensus ranking) according to the axiomatic approach presented
by Cook et al. (2007). This approach minimises the number of violations
between all candidate consensus rankings and all input (partial) rankings,
and draws on a branch and bound algorithm and a heuristic algorithm to
drastically improve speed. The package also provides an option to bootstrap
a consensus ranking based on resampling input rankings (with
replacement). Input rankings can be either incomplete (partial) or complete.
Reference: Cook, W.D., Golany, B., Penn, M. and Raviv, T. (2007) <doi:10.1016/j.cor.2005.05.030>.

Jay Burns

RankAggregator

Aggregation of (Partial) Ordinal Rankings

Adam Butler

lowerBound function

a <code>matrix</code>, which is either the output of <code><a rd-options='' href='evaluationMatrix'>evaluationMatrix</a></code>,
or a subset of the output of <code><a rd-options='' href='evaluationMatrix'>evaluationMatrix</a></code>.

This function is called by <code><a rd-options='' href='RankAggregator'>RankAggregator</a></code>::<code><a rd-options='' href='consensusRanking'>consensusRanking</a></code>.
The lower bound is the absolute lowest value a complete candidate ranking
could attain. Note, this value is not always achievable, so may differ
from the value returned by <code><a rd-options='' href='upperBound'>upperBound</a></code>.<pre> For each pair of \code{Item}s, there are three possible calculations, depending
 on whether both \code{Item}s are in the \code{partial.ranking}, one is in
 and the other is out the \code{partial.ranking}, or both are not in
 the \code{partial.ranking}.
</pre>

lowerBound: Lower bound value

Description

Usage

Arguments

Value