https://github.com/cran/RecordLinkage
Tip revision: 2b541c58b0e09eecd3e8c7c35fdadf24c4b468c4 authored by Andreas Borg on 23 March 2010, 00:00:00 UTC
version 0.2-0
version 0.2-0
Tip revision: 2b541c5
epiClassify.Rd
\name{epiClassify}
\Rdversion{1.1}
\alias{epiClassify}
\title{
Classify record pairs with EpiLink weights
}
\description{
Classifies record pairs as link, non-link or possible link based on
weights computed by \code{\link{epiWeights}} and the thresholds
passed as arguments.
}
\usage{
epiClassify(rpairs, threshold.upper, threshold.lower = threshold.upper)
}
\arguments{
\item{rpairs}{
\code{\link{RecLinkData}} object. Record pairs to be classified.}
\item{threshold.upper}{A numeric value between 0 and 1. Threshold for links.}
\item{threshold.lower}{A numeric value between 0 and 1. Threshold for
possible links.}
}
\details{
The object \code{rpairs} must have weights calculated by \code{\link{epiWeights}},
stored in component \code{Wdata}. All record pairs with weights greater or
equal \code{threshold.upper} are classified as links. Record pairs with
weights smaller than \code{threshold.upper} and greater or equal
\code{threshold.lower} are classified as possible links. All remaining
records are classified as non-links.
}
\value{
A \code{\link{RecLinkResult}} object with classification results stored in
component \code{prediction}, otherwise a copy of \code{rpairs}.
}
\author{
Andreas Borg
}
\seealso{
\code{\link{epiWeights}}
}
\examples{
# generate record pairs
data(RLdata500)
p=compare.dedup(RLdata500,strcmp=TRUE ,strcmpfun=levenshteinSim,
identity=identity.RLdata500)
# calculate weights
p=epiWeights(p)
# classify and show results
summary(epiClassify(p,0.6))
}
\keyword{classif}