StratifiedSampling: stratified sampling

In elqumsan/RNAseqMVA: Application of Multivariate Analysis methods to RNA-seq Data

Description

Perform a stratified sampling from a vector of group labels: each initial group is replaced by random labels proportional to the initial group proportions.

Usage

StratifiedSampling(labels, size = length(labels))

Arguments

labels	a vector with labels to be resampled
size=length(labels)	size of the output (resampled) vector

Author(s)

Jacques.van-Helden\@univ-amu.fr

Examples

## Generate a vector of group labels
labels <- rep(x = c("A", "B", "C", "D"), times = c(256, 64, 16, 4))

## Non-stratified sampling
nstr <- sample(labels, size = length(labels) , replace = FALSE)
table(labels, nstr)
## Repeat this several times: the proportions vary between iterations

## Stratified permutation of group labels
str1 <- StratifiedSampling(labels)
## str1 has same proportions as labels (except for rounding of group sizes)
table(labels, str1)

## Second stratified permutation
str2 <- StratifiedSampling(labels)
## str2 has same proportions as labels (except for rounding of group sizes)
table(labels, str2)

## str1 and str2 are different and independent
table(str1, str2)

## Subsampling
str3 <- StratifiedSampling(labels, size = length(labels) * 2/3)
## str3 has same proportions as labels (except for rounding of group sizes)
table(str3)

elqumsan/RNAseqMVA documentation built on March 10, 2021, 8:10 a.m.