compositions (version 1.40-5)

zeroreplace: Zero-replacement routine

Description

A function to automatically replace rounded zeroes/BDLs in a composition.

Usage

zeroreplace(x,d=NULL,a=2/3)

Arguments

x

composition or dataset of compositions

d

vector containing the detection limits of each part

a

fraction of the detection limit to be used in replacement

Value

an object of the same class as x, where all WZERO values have been replaced. Output contains a further attribute (named Losts), with a logical array of the same dimensions as x, showing which elements were replaced (TRUE) and which were kept unchanged (FALSE).

Details

If d is given, zeroes from each column of x are replaced by the corresponding detection limit contained there, scaled down by the value of a (usually a scalar, although if it is a vector it will be recycled with a warning). The variable d should be a vector of length equal to ncol(x) or a matrix of the same shape as x.

If d=NULL, then the detection limit is extracted from the data set, if it is available there (i.e., if there are negative numbers). If no negative number is present in the data set, and no value is given for d, the result will be equal to x. See compositions.missings for more details on the missing policy.

References

Aitchison, J. (1986) The Statistical Analysis of Compositional Data Monographs on Statistics and Applied Probability. Chapman & Hall Ltd., London (UK). 416p.

Mart\'in-Fern\'andez, J.A.; Barcel\'o-Vidal, C. and Pawlowsky-Glahn, V. (2003) Dealing With Zeros and Missing Values in Compositional Data Sets Using Nonparametric Imputation. Mathematical Geology, 35 , 253-278

http://ima.udg.es/Activitats/CoDaWork03

http://ima.udg.es/Activitats/CoDaWork05

See Also

compositions.missings,getDetectionlimit

Examples

# NOT RUN {
data(SimulatedAmounts)
x <- acomp(sa.lognormals)
xnew <- simulateMissings(x,dl=0.05,knownlimit=FALSE)
xnew
xrep <- zeroreplace(xnew,0.05)
xrep
# }