CODEWITHSUNDEEP
rfunctiondistribution
user4574145
user4574145

Reputation:

Looping Multiple Variable Values

I am trying to run a simulation that creates two independent normal distributions from different n-counts, means, and standard deviations. For example, n-counts from 100 to 1000, means from 0 to 100, and standard deviations from 1 to 10, where each element from each vector combines with every other element of the other two vectors, like so: 100, 0, 1; 100, 0, 2; 100, 0, 3;...100, 1, 1; 100, 1, 2; 100, 1, 3;...999, 0, 1; etc.

I first tried to create three vectors that had these numbers:

    ncount <- 100:1000
    means <- 0:100
    std <- 1:10

I also wrote a rudimentary function like so:

    distr <- function(count, mean, stddev) {
      distr1 <- rnorm(count1, mean1, stddev1)
      hist(distr1)
    }

At this point, I'm stuck (mostly because of not knowing) what R can do to cycle through each of the values in these vectors to produce the different distributions. Suggestions?

Upvotes: 1

Views: 36

Answers (2)

Jthorpe
Jthorpe

Reputation: 10167

While @Marat Talipov's answer is ideal from many respects, don't forget that you can also use nested for loops, as in:

for(ncount in 100:1000)
for(means in 0:100)
for(std in 1:10)
    distr(ncount,means,std)

Upvotes: 0

Marat Talipov
Marat Talipov

Reputation: 13304

You can use expand.grid, which will create all combinations of ncount/means/std, and Vectorize, which will 'teach' your function to treat input arguments in a row-wise manner:

ncount <- 100:101
means <- 0:2
std <- 1:3
distr <- function(count, mean, stddev) {
    distr1 <- rnorm(count, mean, stddev)
    hist(distr1)
}

d <- expand.grid(ncount=ncount,means=means,std=std)

Vectorize(distr)(d$ncount,d$means,d$std)

Upvotes: 1

Related Questions