% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/MixMC.R
\name{normalise.clr.within.features}
\alias{normalise.clr.within.features}
\title{Apply centered log-ratio normalisation within features only}
\usage{
normalise.clr.within.features(input, offset = 0)
}
\arguments{
\item{input}{Data as proportions 0-1.}

\item{offset}{Optional offset to apply to raw data to avoid logging of zero
values. Only needed if any zeroes are present - should generally be set very
small, e.g. 0.000001.}
}
\value{
Data normalised by the within-feature CLR method
}
\description{
Apply centered log-ratio (CLR) normalisation to other compositional data, but
restrict normalisation to *within* features only. This is another method for
converting the compositional data, i.e. proportional data in the range 0..1
to Euclidean space which is most appropriate for the linear models. The
implementation is the same as CLR, but on transposed input data (which is
then transposed back to the input orientation). Note this is experimental,
though it does give a sensible normalisation.
}
\examples{
\dontrun{
normalise.clr.within.features(data.proportional)
}
}