colCenterScale: colCenterScale

In pipeComp: pipeComp pipeline benchmarking framework

Description

Matrix scaling by centering columns separately and then performing variance scaling on the whole matrix, in a NA-robust fashion. With the default arguments, the output will be the number of (matrix-)median absolute deviations from the column-median.

Usage

colCenterScale(
  x,
  centerFn = median,
  scaleFn = function(x, na.rm) median(abs(x), na.rm = na.rm)
)

Arguments

x	A numeric matrix.
centerFn	The function for calculating centers. Should accept the 'na.rm' argument. E.g. 'centerFn=mean' or 'centerFn=median'.
scaleFn	The function for calculating the (matrix-wise) scaling factor. Should accept the 'na.rm' argument. Default 'median(abs(x))'.

Value

A scaled matrix of the same dimensions as 'x'.

Examples

# random data with column mean differences
d <- cbind(A=rnorm(5, 10, 2), B=rnorm(5, 20, 2), C=rnorm(5,30, 2))
colCenterScale(d)

pipeComp documentation built on Nov. 8, 2020, 7:35 p.m.