compositions (version 2.0-2)

sumprojector: Compute the global projector to the observed subspace.

Description

Routines to compute the global projector to the observed subspace, down-weighting the subspaces with more missing values.

Usage

sumMissingProjector(x,...)
# S3 method for acomp
sumMissingProjector(x,has=is.NMV(x),...)
# S3 method for aplus
sumMissingProjector(x,has=is.NMV(x),...)
# S3 method for rcomp
sumMissingProjector(x,has=!(is.MAR(x)|is.MNAR(x)),...)
# S3 method for rplus
sumMissingProjector(x,has=!(is.MAR(x)|is.MNAR(x)),...)
# S3 method for rmult
sumMissingProjector(x,has=is.finite(x),...)

Arguments

x

a dataset of some type containing missings

has

the values to be regarded as non missing

…

further generic arguments that might be useful for other functions.

Value

The matrix of rotation/re-weighting of the original data set, down-weighting the subspaces with more missing values. This matrix is useful to obtain estimates of the mean (and variance, in the future) still unbiased in the presence of lost values (only of type MAR, stricly-speaking, but anyway useful for any type of missing value, when used with care). This matrix is the Fisher Information in the presence of missing values.

Missing Policy

No missing policy is given by the routine itself. Its treatment of missing values depends on the "has" argument.

Details

The function missingProjector generates a list of N square matrices of dimension DxD (with N and D respectively equal to the number of rows and columns in x). Each of these matrices gives the projection of a data row onto its observed sub-space. Then, the function sumMissingProjector takes all these matrices and sums them in a efficient way, generating a "summary" of observed sub-spaces.

References

Boogaart, K.G. v.d., R. Tolosana-Delgado, M. Bren (2006) Concepts for handling of zeros and missing values in compositional data, in E. Pirard (ed.) (2006)Proccedings of the IAMG'2006 Annual Conference on "Quantitative Geology from multiple sources", September 2006, Liege, Belgium, S07-01, 4pages, http://stat.boogaart.de/Publications/iamg06_s07_01.pdf

See Also

missingProjector, clr,rcomp, aplus, princomp.acomp, plot.acomp, boxplot.acomp, barplot.acomp, mean.acomp, var.acomp, variation.acomp, cov.acomp, msd

Examples

# NOT RUN {
data(SimulatedAmounts)
sumMissingProjector(acomp(sa.lognormals))
sumMissingProjector(acomp(sa.tnormals))
# }