simboil

The goal of simboil is to stream line simulation workflows. It hopes to bring simmering simmering simmering simulations to a boil.

Installation

You can install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("mncube/simboil")

Use Monte Carlo simulation to test a condition

The default setting of simtest simulates a coin flip and tests whether 10 flips has 3 or more head (where head is defined as 1). The function and example was inspired by A Beginner’s Guide to Monte Carlo Simulations presented by Dan Uehara for UseR Oslo: https://www.youtube.com/watch?v=g2Uu9m0IlFE

library(simboil)
coinflips <- simtest()
mean(coinflips)
#> [1] 0.98

Use simulation to get quantiles

In the example below, simq is used to compare a practice test to the test mean and standard deviation from the past 3 years

#Test scores from the new test administered over the past 3 years is not 
#available but the mean (75) and sd (14) is known
sim_scores <- as.vector(rnorm(1000, 75, 14))

#Practice test scores for low performing math class 
practice_test <- as.vector(rnorm(30, 68, 7))

#See how low performers compare to past performers

#Run simulation
getqs <- simq(frame = sim_scores, ref_samp = practice_test, 
              main = "Distribution of Past Tests", xlab = "Sample Means")

#Get requested quantiles (example uses defaults of lowp = .025, highp = .975)
getqs$quantiles
#>     2.5%    97.5% 
#> 70.37517 80.40910

#Compare practice test to distribution of past tests
getqs$histogram
#> $breaks
#>  [1] 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
#> 
#> $counts
#>  [1]    4    3   19   53  113  249  502  863 1171 1514 1565 1360 1096  709  416
#> [16]  228   89   33    8    3    1    1
#> 
#> $density
#>  [1] 0.0004 0.0003 0.0019 0.0053 0.0113 0.0249 0.0502 0.0863 0.1171 0.1514
#> [11] 0.1565 0.1360 0.1096 0.0709 0.0416 0.0228 0.0089 0.0033 0.0008 0.0003
#> [21] 0.0001 0.0001
#> 
#> $mids
#>  [1] 65.5 66.5 67.5 68.5 69.5 70.5 71.5 72.5 73.5 74.5 75.5 76.5 77.5 78.5 79.5
#> [16] 80.5 81.5 82.5 83.5 84.5 85.5 86.5
#> 
#> $xname
#> [1] "samps"
#> 
#> $equidist
#> [1] TRUE
#> 
#> attr(,"class")
#> [1] "histogram"

Compare the empirical cdf to the true cdf using the plots

Plot the empirical vs true cdf for the uniform distribution as follows:

evt_cdf(ylab = "Cumulative Probability", col = "blue", main = "Uniform ECDF (black) vs Uniform CDF (blue)")

To get a more accurate ECDF plot, increase the number of samples using n

evt_cdf(n = 10^5, ylab = "Cumulative Probability", col = "blue", main = "Uniform ECDF (black) vs Uniform CDF (blue)")

You can also change from the uniform distribution to other distributions such as the normal distribution by using anonymous functions

evt_cdf(n = 200, ylab = "Cumulative Probability", col = "blue", main = "Normal ECDF (black) vs Normal CDF (blue)",
        remp = function(x)stats::rnorm(x),
        ptrue = function(x)stats::pnorm(x))