Is the a way to regenerate data for 500 times?-CodePudding

library(MASS)

# set seed and create data vectors
#set.seed(98989) <- for replicating results of betas in 1-2 1-3
sample_size <- 200                                       
sample_meanvector <- c(3, 4)                                   
sample_covariance_matrix <- matrix(c(2, 1, 1, 2),
                                   ncol = 2)

# create bivariate normal distribution
sample_distribution <- mvrnorm(n = sample_size,
                               mu = sample_meanvector, 
                               Sigma = sample_covariance_matrix)
#Convert the datatype
df_sample_distribution <- as.data.frame(sample_distribution)

Is there a way to put this entire chunk of code in a loop and regenerate it for 500 times? Would be even better if i can store them somewhere.

CodePudding user response：

You might use replicate()

library(MASS)
out <- replicate(3, simplify = FALSE, {sample_size <- 200                                       
          sample_meanvector <- c(3, 4)                                   
          sample_covariance_matrix <- matrix(c(2, 1, 1, 2),
                                             ncol = 2)
          
          # create bivariate normal distribution
          sample_distribution <- mvrnorm(n = sample_size,
                                         mu = sample_meanvector, 
                                         Sigma = sample_covariance_matrix)
          #Convert the datatype
          df_sample_distribution <- as.data.frame(sample_distribution)

          head(df_sample_distribution) # for shorter output
          })

out
#> [[1]]
#>         V1       V2
#> 1 3.195478 4.393699
#> 2 2.553590 5.065685
#> 3 2.822811 2.389559
#> 4 2.267116 4.076016
#> 5 1.659459 3.830608
#> 6 1.377554 4.009023
#> 
#> [[2]]
#>           V1       V2
#> 1  2.8850139 3.107203
#> 2  3.0313680 5.163229
#> 3  3.8649482 4.594017
#> 4  3.2747060 4.085805
#> 5 -0.1640264 3.628542
#> 6  3.6504855 4.747372
#> 
#> [[3]]
#>          V1       V2
#> 1 1.3230817 4.075396
#> 2 3.6049470 6.293968
#> 3 6.1211276 7.673592
#> 4 5.2955379 6.736665
#> 5 0.9032304 2.606501
#> 6 3.6034566 3.880563

^{Created on 2022-12-04 with reprex v2.0.2}

CodePudding user response：

Yes, you can put the code in a loop and generate the sample data 500 times. Here is an example of how you can do that:

# set the number of iterations
num_iterations <- 500

# create an empty list to store the generated data
generated_data <- list()

# loop through the number of iterations
for (i in 1:num_iterations) {
  # set the seed
  set.seed(i)
  
  # create the sample data using the mvrnorm function
  sample_distribution <- mvrnorm(n = sample_size,
                                 mu = sample_meanvector, 
                                 Sigma = sample_covariance_matrix)
  
  # convert the data to a data frame
  df_sample_distribution <- as.data.frame(sample_distribution)
  
  # store the generated data in the list
  generated_data[[i]] <- df_sample_distribution
}

# you can access the generated data using the list index, for example:
generated_data[[1]] # will return the first generated data

You can also store the generated data in a data frame by using the rbind function to combine the data frames in the list into a single data frame. Here is an example of how you can do that:

# create an empty data frame to store the generated data
generated_data_df <- data.frame()

# loop through the generated data list
for (i in 1:num_iterations) {
  # bind the data frame at the current index to the generated data data frame
  generated_data_df <- rbind(generated_data_df, generated_data[[i]])
}

# generated_data_df will now contain all the generated data

Alternatively, you can use the do.call and rbind functions to combine the data frames in the list into a single data frame in a single step, like this:

# create the data frame using the do.call and rbind functions
generated_data_df <- do.call(rbind, generated_data)

CodePudding user response：

Yes, you can use a for loop to generate the data multiple times. Here is an example:

# set seed and create data vectors
set.seed(98989)
sample_size <- 200                                       
sample_meanvector <- c(3, 4)                                   
sample_covariance_matrix <- matrix(c(2, 1, 1, 2),
                                   ncol = 2)

# create a list to store the data frames
df_list <- list()

# loop to generate the data
for (i in 1:500) {
  # create bivariate normal distribution
  sample_distribution <- mvrnorm(n = sample_size,
                                 mu = sample_meanvector, 
                                 Sigma = sample_covariance_matrix)
  # Convert the data type
  df_sample_distribution <- as.data.frame(sample_distribution)
  # add the data frame to the list
  df_list[[i]] <- df_sample_distribution
}

This code will generate 500 data frames, each containing the bivariate normal distribution data. The data frames will be stored in the df_list list. You can access each data frame by indexing the list, for example df_list[[1]] will give you the first data frame.