Varren Varren - 2 months ago 6
R Question

Can I have a vector of vectors in R

I want to produce all possible

w
when
s
range from -2 to 2 by 0.1, where
w_min
and
w_max
are 5 x 1 vectors. But I don't know how to represent the result. I think I need a vector with each element as a vector, too, i.e., a vector of vectors.

s <- seq(-2, 2, by = 0.1)

result = c()

for (i in 1:20) {
w = s[i] * w_min + (1 - s[i]) * w_max
## what do I need to do here??
}

result

Answer

You want a matrix, where you have many columns while each column is a vector.


To provide a toy example, I need to make your "w_min and w_max are 5 * 1 vectors" concrete:

## note, they are just plain vectors without dimension
## if you have a `5 * 1` matrix, use `c(w_min)` and `c(w_max)` to drop dimension
w_min <- 1:5
w_max <- 2:6

Also, to make the example small, I will consider s <- seq(-2, 2, by = 1) with step 1.

First, consider the loop-based method:

w <- matrix(0, 5, length(s))  ## set up a `5 * length(s)` matrix
for (i in 1:length(s)) {
  ## fill i-th column of the matrix
  w[, i] <- s[i] * w_min + (1 - s[i]) * w_max
  }

w
#     [,1] [,2] [,3] [,4] [,5]
#[1,]    4    3    2    1    0
#[2,]    5    4    3    2    1
#[3,]    6    5    4    3    2
#[4,]    7    6    5    4    3
#[5,]    8    7    6    5    4

Then, the vectorized method:

## read `?outer`; the default function to apply is `FUN = "*"` for multiplication
w <- outer(w_min, s) + outer(w_max, 1 - s)

w
#     [,1] [,2] [,3] [,4] [,5]
#[1,]    4    3    2    1    0
#[2,]    5    4    3    2    1
#[3,]    6    5    4    3    2
#[4,]    7    6    5    4    3
#[5,]    8    7    6    5    4

Apart from a matrix, you can also store your result in a list of vectors.

w <- vector("list", length(s))  ## set up a `length(s)` list
for (i in 1:length(s)) {
  ## fill i-th element of the list; note the `[[i]]`
  w[[i]] <- s[i] * w_min + (1 - s[i]) * w_max
  }

w
#[[1]]
#[1] 4 5 6 7 8
#
#[[2]]
#[1] 3 4 5 6 7
#
#[[3]]
#[1] 2 3 4 5 6
#
#[[4]]
#[1] 1 2 3 4 5
#
#[[5]]
#[1] 0 1 2 3 4

But there is no real vectorization approach here. We can at most hide the loop by an lapply:

w <- lapply(s, function (x) x * w_min + (1 - x) * w_max)

w
#[[1]]
#[1] 4 5 6 7 8
#
#[[2]]
#[1] 3 4 5 6 7
#
#[[3]]
#[1] 2 3 4 5 6
#
#[[4]]
#[1] 1 2 3 4 5
#
#[[5]]
#[1] 0 1 2 3 4