This funciton calls RankAggregator::consensusRanking to
return a best consensus (or median) ranking for a set of (partial) rankings.
This function also provides an optional bootstrap resampling procedure to
give user-defined confidence intervals and average rank positions with
the consensus ranking.
consensusRankingBoot(
x,
bootstrap,
nboot = 10000,
conf.int = 0.95,
prog.upd = TRUE
)a data.frame containing columns titled Reviewer, Item,
Ranking. On data structure, Reviewer and Item must be
character, and Ranking must be numeric. Each row of x
identifes the rank position that a single Reviewer awarded a single
Item
a logical value indicating whether to bootstrap the rank aggregation procedure.
a numeric value for bootstrap replicates. Default value is 10000.
a numeric value >0 and <1. Default value is 0.95, which sets
confidence interval at 95% level.
a logical value indicating whether the user wants progress updates on the bootstrap procedure.
If bootstrap is FALSE, a data.frame is returned,
with two columns: Item and Rank.est, where Item
is a Factor containing all unique Item's from the input
data.frame x, and where Rank.est is the estimated
(numeric) rank position based on the consensusRanking() rank
aggregation procedure.#'
If bootstrap is TRUE, a list is returned, with two
elements:
$summaryTable is a data.frame with six columns: Item
Rank.est, Rank.cilo, Rank.cihi, Rank.median,
Rank.mean. Where Item and Rank.est are as described
above, Rank.cilo and Rank.cihi are the estimates for the
low and high confidence intervals, respectively. Rank.median and
Rank.mean both describe the average rank positions.
$bootstrapData is an array containing estimated (numeric)
rank positions based on the consensusRanking() rank aggregation
procedure with resampled data. NA denotes estimated rankings that
were discarded due to not containing all Items.
Cook, W.D., Golany, B., Penn, M. and Raviv, T., 2007. Creating a consensus ranking of proposals from reviewers partial ordinal rankings. Computers & Operations Research, 34, pp.954-965. Marshall, E.C., Sanderson, C., Spiegelhalter, D.J. and McKee, M., 1998. Reliability of league tables of in vitro fertilisation clinics: retrospective analysis of live birth ratesCommentary: How robust are rankings? The implications of confidence intervals. Bmj, 316, pp.1701-1705.
Calls the internal function consensusRanking, which calls the other
internal functions evaluationMatrix, consensusRanking,
extendRanking, lowerBound,
upperBound