For loop within a function using function arguments

Question

I am running regressions on a data set with > 4 million observations. I am trying to do a sensitivity analysis to see how sensitive my regression coefficients are to different sample sizes. I want to:

Take a random sample of n data points
Run a regression on the sample
Extract the coefficient

I want to replicate the above steps 400 times to create a distribution of coefficients and plot the mean and confidence intervals. I have created two functions to do this. The first, coef_sampler, runs a regression on a sample of the data and returns the coefficient:

coef_sampler <- function(data,n){
      model <- lm(y ~ x, data = data[sample(nrow(data), n),])
      return(model[["coefficients"]][[1]])
  }

The second, my_function, replicates this process 400 times for different sample sizes, generating vectors of coefficients for different sample sizes. It then calculates the mean of each vector along with it's upper and lower bounds using the CI() function from Rmisc:

my_function <- function(data, n, nrep, to, by){

  for (j in seq.int(n, to, by)){
    plist <- data.frame(pval = replicate(nrep, coef_sampler(data,j)))

    if(!exists("ci_mat")) {
      ci_mat <- data.frame(CI(plist[,1]))
    } else {
      ci_mat <- cbind(ci_mat, data.frame(CI(plist[,1])))
    }
  }
}

When I just run the for loop with the values I want for number of repetitions, and a sequence of sample sizes, it works fine. When I call my_function() with the above values, it simply doesn't work. No data frame of means and confidence intervals is generated. Note that I call the coef_sampler() function within my_function, could this be causing the problem? Or is it using a for loop that uses the function arguments causing the problem.

Hopefully my question makes sense, any and all help would be greatly appreciated!

CI is from Rmisc. It just calculates confidence intervals and returns the upper bound, mean, and lower bound — tsgsteele
– tsgsteele, Commented Nov 17, 2019 at 22:00
Just a general reminder: for a reproducible minimal example, please include all needed packages along with some sample data, e.g. with dput() — mnist
– mnist, Commented Nov 17, 2019 at 22:01

mnist · Accepted Answer · 2019-11-17 22:04:29Z

2

Just include a return at the end of your function

library(Rmisc)

my_data <- tibble(x = 1:5,
                  y = rnorm(5))

coef_sampler <- function(data,n){
  model <- lm(y ~ x, data = data[sample(nrow(data),n),])
  return(model[["coefficients"]][[1]])
}

my_function <- function(data,n,nrep,to,by){
  for (j in seq.int(n,to,by)){
    plist <- data.frame(pval = replicate(nrep,coef_sampler(data,j)))

    if(!exists("ci_mat")){
      ci_mat <- data.frame(CI(plist[,1]))
    } else {
      ci_mat <- cbind(ci_mat, data.frame(CI(plist[,1])))
    }
  }
  return(ci_mat)
}

my_function(my_data, 1, 2, 5, 1)

answered Nov 17, 2019 at 22:04

mnist

7,0242 gold badges20 silver badges44 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

tsgsteele Over a year ago

Will upvote this as well because you basically posted at the exact same time. Thank you!

slava-kohut · Accepted Answer · 2019-11-17 22:04:07Z

1

The problem is that your function does not return the populated ci_mat. This variables only lives in the function environment and gets destroyed after the function run:

df <- data.frame(x = 1:100, y = 4*(1:100) + 5 + rnorm(100,5,10))

coef_sampler <- function(data,n){
  model <- lm(y ~ x, data = data[sample(nrow(data),n),])
  return(model[["coefficients"]][[1]])
}

my_function <- function(data,n,nrep,to,by){

  for (j in seq.int(n,to,by)){
    plist <- data.frame(pval = replicate(nrep,coef_sampler(data,j)))

    if(!exists("ci_mat")){
      ci_mat <- data.frame(Rmisc::CI(plist[,1]))
    }else{
      ci_mat <- cbind(ci_mat, data.frame(Rmisc::CI(plist[,1])))
    }
  }
  ci_mat
}    

# sample call
my_function(df, 5, 10, 100,1)

Make your function return ci_mat.

answered Nov 17, 2019 at 22:04

slava-kohut

4,2331 gold badge10 silver badges24 bronze badges

1 Comment

tsgsteele Over a year ago

Wow. I probably should have seen that. Works great now, I love stackoverflow! Thank you!

Collectives™ on Stack Overflow

For loop within a function using function arguments

2 Answers 2

1 Comment

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related