I have been reading permutation test in Wikipedia. There is no mention about stratified permutation. Is there something like stratified permutation test?
x <- rbind(matrix(rnorm(100, mean = 0, sd = 0.5), ncol = 2), matrix(rnorm(100, mean = 1, sd = 0.5), ncol = 2), matrix(rnorm(100, mean = 2, sd = 0.5), ncol = 2)) y <- c(rep(0,50), rep(1,50), rep(2,50)) data <- data.frame(x,y) perms <- factorial(nrow(data)) # 5.713384e+262
Is it advisable to do permutations
5.713384e+262) or only
nrow**different possible labels(
150**3=3375000), if all possible permutations need to be considered?
The observed data is shuffled by assigning different outcome values to each observation from the set of observed outcomes right? If yes, I have a few questions:
- Should the assignment be always with the same frequency? As in the above example, the permuted
yvalues should always be having same frequency of 50?