Suppose I have a list like this in R:
L <- list(
c(0.68, 0.78, 0.63, 0.2, 0.89, 0.81),
c(0.93, 0.89, 0.77, 0.86, 0.49, 0.2),
c(0.88, 0.08, 0.15, 0.35, 0.6, 0.45),
c(0.07, 0.85, 0.49, 0.92, 0.87, 0.34)
)
L
[[1]]
[1] 0.68 0.78 0.63 0.20 0.89 0.81
[[2]]
[1] 0.93 0.89 0.77 0.86 0.49 0.20
[[3]]
[1] 0.88 0.08 0.15 0.35 0.60 0.45
[[4]]
[1] 0.07 0.85 0.49 0.92 0.87 0.34
And I want to find the correlation coefficient between each pair of vectors but over multiple lengths, e.g, cor(L[[i]][1:t],L[[j]][1:t]) that t changes from 2 to 6. Is there any solution to perform it with the apply family function?
We may use combn
combn(L, 2, FUN = function(x) lapply(2:6, function(t)
cor(x[[1]][seq_len(t)], x[[2]][seq_len(t)])), simplify = FALSE)
If we need a matrix output
combn(L, 2, FUN = function(x) do.call(c, lapply(2:6,
function(t) cor(x[[1]][seq_len(t)], x[[2]][seq_len(t)]))))
-output
[,1] [,2] [,3] [,4] [,5] [,6]
[1,] -1.0000000 -1.00000000 1.0000000 1.00000000 -1.00000000 -1.0000000
[2,] 0.5765567 -0.26596468 0.6204701 0.63428533 -0.28302468 -0.9210074
[3,] 0.1641508 -0.03795219 -0.4358611 0.63452989 -0.24528496 -0.7680795
[4,] -0.4623420 0.14114454 -0.1512739 -0.09921710 -0.41171957 -0.5675466
[5,] -0.4846004 0.15007699 -0.2377249 -0.09824801 0.08412466 -0.5485821