clr: clr The centered log-ratio transformation

In zdk123/compPLS: Conduct Partial Least Squares regression for compositional data

clr The centered log-ratio transformation

Description

The centered log-ratio transformation for compositional data (not necessarily closed/normalized!)

Usage

clr(x, ...)

## Default S3 method:
clr(x, base = exp(1), tol = .Machine$double.eps)

## S3 method for class 'matrix'
clr(x, mar = 2, base = exp(1), tol = .Machine$double.eps)

## S3 method for class 'data.frame'
clr(x, mar = 2, ...)

Arguments

x	a numeric data vector containing components of a composition
...	additional arguments
base	base of log to use, default is natural log
tol	machine tolerance for a zero count, default is machine tol (.Machine$double.eps)

Details

The clr is computed as x[i] = log (x[i] / exp(mean(log(x))))

Value

clr transformed x

Examples

# vector examples:
clr(norm_to_total(1:10))
clr(1:10)

# matrix examples:
dmat <- matrix(exp(rnorm(110)), 10)
clr(dmat, 1)
clr(dmat, 2)

zdk123/compPLS documentation built on April 24, 2022, 2:44 p.m.