weighted_ilr: Weighted ILR Transform

In jsilve24/philr: Phylogenetic partitioning based ILR transform for metagenomics data

Weighted ILR Transform

Description

Calculated using weighted CLR transform (clrp)

Usage

ilrp(y, p, V)

ilrpInv(y.star, V)

Arguments

y	shifted data matrix (e.g., output of shiftp)
p	weights (should not be closed)
V	weighted contrast matrix (e.g., output of buildilrBasep)
y.star	a data matrix that represents data transformed by ilrp

Value

matrix

Author(s)

Justin Silverman

References

J. J. Egozcue, V. Pawlowsky-Glahn (2016) Changing the Reference Measure in the Simplex and its Weighting Effects. Austrian Journal of Statistics 45(4):25-44

See Also

philrInv

Examples

# Weights
p <- seq(.1,1,by=.2)

# Shifted Composition
c <- t(rmultinom(10,100,c(.1,.6,.2,.3,.2))) + 0.65   # add a small pseudocount
x <- miniclo(c)
y <- shiftp(x, p)

# Contrast Matrix
tr <- named_rtree(5)
sbp <- phylo2sbp(tr)
V <- buildilrBasep(sbp, p)

y.star <- ilrp(y, p, V)
y.star

# Untransform data (note use of shiftp and miniclo to return to x)
y.closed <- ilrpInv(y.star, V)
all.equal(miniclo(shiftpInv(y.closed, p)), x, check.attributes=FALSE)

jsilve24/philr documentation built on Oct. 17, 2024, 4:59 p.m.