Type: Package
Title: Parallel Simulation Studies
Version: 0.1.5
Author: Sacha Epskamp
Maintainer: Sacha Epskamp <mail@sachaepskamp.com>
Description: Perform flexible simulation studies using one or multiple computer cores. The package is set up to be usable on high-performance clusters in addition to being run locally, see examples on https://github.com/SachaEpskamp/parSim.
Imports: dplyr, methods, pbapply, snow, data.table, utils
Suggests: ggplot2, tidyr
License: GPL-2
NeedsCompilation: no
Packaged: 2023-05-16 11:50:26 UTC; sachaepskamp
Repository: CRAN
Date/Publication: 2023-05-16 12:10:02 UTC

Parallel Simulator

Description

Takes a set of conditions and an R expression and returns a data frame with simulation results. parSim is based on dplyr functions, and if you want to use the data.table package to make your simulation a bit faster, use parSim_dt. See details.

Usage

parSim(..., expression, reps = 1, write = FALSE, name, 
       nCores = 1, export, exclude, debug = FALSE, 
       progressbar = TRUE)
                   
parSim_dt(..., expression, reps = 1, write = FALSE, name, 
          nCores = 1, export, exclude, debug = FALSE, 
          progressbar = TRUE)

Arguments

...

Vectors indicating any number of conditions. For example, if you want to vary sample size between N = 100, 250, and 1000, and a regression slope between beta = 0, 0.5, and 1, you can assign as first two arguments sampleSize = c(100, 250, 1000), beta = c(0, 0.5, 1).

expression

An R expression that uses the conditions as object names. For example, if the conditions sampleSize = c(100, 250, 1000) is used, then in the R expression you can use the object sampleSize, which may be 100, 250 or 1000 depending on the simulation condition.

reps

Number of times each condition has to be replicated.

write

Logical, should the results be written to a file instead of returned as a dataframe?

name

Name of the file if write = TRUE

nCores

Number of cores to use. NOTE: Only setting nCores to 1 allows you to see a progress bar and to use browser() in the R expression for debugging.

export

A character string of objects to be exported. Only needed if nCores > 1.

exclude

A list with logical calls to exclude cases. Written as formula.

debug

Allows for some debugging controls and output. Not recommended.

progressbar

Logical: should a progress bar be shown. Setting this to FALSE will make simulations much faster!

Details

The R expression should use object names assigned as conditions, and should return a list with single values, or a data frame / data table. If you want to output more than one row of results per condition, you may return a data frame / data table with multiple rows. When using multiple cores, note that all packages should be loaded in the R expression, all objects needed should be exported using the export object, and you will not see a progress bar.

Value

parSim outputs a data frame with the results of every iteration as a row.

parSim_dt outputs a data table with the results of every iteration as a row.

Author(s)

Sacha Epskamp <mail@sachaepskamp.com>

Xinkai Du <xinkai.du.xd@gmail.com>

Examples

# Some function we might use:
bias <- function(x,y){abs(x-y)}

# Run the simulation:
Results <- parSim(
  # Any number of conditions:
  sampleSize = c(50, 100, 250),
  beta = c(0, 0.5, 1),
  sigma = c(0.25, 0.5, 1),
  
  # Number of repititions?
  reps = 25, # more is better!
  
  # Parallel?
  nCores = 1,
  
  # Write to file?
  write = FALSE,
  
  # Export objects (only needed when nCores > 1):
  export = c("bias"),
  
  # R expression:
  expression = {
    # Load all R packages in the expression if needed
    # library(...)
    
    # Want to debug? Enter browser() and run the function. Only works with nCores = 1!
    # browser()
    
    # Enter whatever codes you want. I can use the conditions as objects.
    X <- rnorm(sampleSize)
    Y <- beta * X + rnorm(sampleSize, sigma)
    fit <- lm(Y ~ X)
    betaEst <- coef(fit)[2]
    Rsquared <- summary(fit)$r.squared
    
    # Make a data frame with one row to return results (multple rows is 
    # also possible but not recommended):
    data.frame(
      betaEst = betaEst,
      bias = bias(beta,betaEst),
      Rsquared = Rsquared
    )
  }
)

# Analyze the results:
library("ggplot2")
library("tidyr")

# We want to plot bias and R-squared. Let's first make it long format:
Long <- gather(Results,metric,value,bias:Rsquared)

# Make factors with nice labels:
Long$sigmaFac <- factor(Long$sigma,levels = c(0.25,0.5,1), 
    labels = c("Sigma: 0.025", "Sigma: 0.5", "Sigma: 1"))

# Now let's plot:
g <- ggplot(Long, aes(x = factor(sampleSize), y = value, fill = factor(beta)))  +
  facet_grid(metric ~ sigmaFac, scales = "free") + 
  geom_boxplot() + 
  theme_bw() + 
  xlab("Sample size") + 
  ylab("") + 
  scale_fill_discrete("Beta")
print(g)