Construct and modify data simulations

Usage

datasim_make(...)

datasim_to_igraph(sim, report_hidden = FALSE)

datasim_intervene(sim, ...)

Arguments

sim: The data simulation object to be modified.
report_hidden: If TRUE, show the hidden nodes (nodes whose names begin with a dot.)
...: Descriptions of the nodes in the simulation, written in assignment form. See details.

Value

an object of class "datasim". Internally, this is a list of the R assignment expressions used when running the simulation.

Details

Simulations in LSTbook are first specified by providing the code for each node (which can be written in terms of the values of other nodes). Once constructed, data can be extracted from the simulation using datasim_run(n) or the generic sample(n).

Each argument defines one node in the simulation. The argument syntax is unusual, using assignment. For instance, an argument y <- 3*x + rnorm(n) defines a node named y. The R code on the RHS of the assignment operator (that is, 3*x + rnorm(n) in the example) will be used by datasim_run() to generate the y variable when the simulation is run. Nodes defined by previous arguments can be used in the code for later arguments.

Helper functions such as mix_with(), categorical(), and several others can be used within the node to perform complex operations.

Remember to use commas to separate the arguments in the normal way.

Examples

Simple_sim <- datasim_make(x <- rnorm(n, sd=2), y <- 3*x + rnorm(n))
Simple_sim |> datasim_run(n = 5)
#> # A tibble: 5 × 2
#>        x      y
#>    <dbl>  <dbl>
#> 1  0.361   1.32
#> 2 -6.55  -18.4 
#> 3 -4.70  -14.9 
#> 4 -4.01  -11.2 
#> 5  1.95    7.43