CODEWITHSUNDEEP
rfor-loopsumresampling
user3919708
user3919708

Reputation: 141

Sample 5 times and sum, with replacement

I am trying to sample with replacement with a slight twist. I want to sample a list, vector, etc... 5 times and sum that. And replace the 5 sampled values, and do this 1000 times.

x<- 1:1000
samp <- matrix(NA, ncol = 1, nrow = 1000)
for(i in length(x)){
samp[i,] <- sum(sample(x,5,replace = TRUE))}

What I cant figure out is why this loop isnt working

Upvotes: 2

Views: 663

Answers (1)

LyzandeR
LyzandeR

Reputation: 37879

You missed 1: in front of length(x) (i.e. from 1 to length of x by 1). Your code should be like this:

x<- 1:1000
samp <- matrix(NA, ncol = 1, nrow = 1000)
for(i in 1:length(x)){
  samp[i,] <- sum(sample(x,5,replace = TRUE))}

Which works great:

> str(samp)
 int [1:1000, 1] 2715 2312 3180 1364 2851 2429 2888 2381 2772 2317 ...

Also, for-loops have a bad reputation of being slow in R so you might want to consider other ways of looping like using the C++ way with (for example) replicate like this:

Define the function:

myfunc <- function(x) {
  sum(sample(x,5,replace = TRUE))
}

And then use it like this:

x <- 1:1000
mymat <- matrix(replicate(1000, myfunc(x)), ncol=1)

> str(mymat)
 int [1:1000, 1] 2481 2236 2492 1759 1905 3243 2606 2624 3013 2309 ...

Upvotes: 2

Related Questions