sim: Simulate data for rosetta

In bklamer/rosetta: Use Latent Traits to Concatenate Imperfectly Matched Datasets

Description

This function simulates 'complete' and 'missing' datasets. For each independent dataset, one variable per domain will be set to missing.

Usage

sim(loading = matrix(c(0.9, 0.8, 0.7, rep(0, 9), 0.6, 0.7, 0.8, rep(0,
  9), 0.8, 0.9, 0.6), ncol = 3), correlation = matrix(c(1, 0.2, 0.4, 0.2,
  1, 0.3, 0.4, 0.3, 1), ncol = 3), factor_structure = list(a = c(1, 2,
  3), b = c(1, 2, 3), c = c(1, 2, 3)), n_rows = 1000, n_datasets = 3,
  seed = NULL)

Arguments

loading	The measurement model for x. The 'fx' argument as found in psych::sim.structure.
correlation	The structure matrix of the latent variables. The 'Phi' argument as found in psych::sim.structure.
factor_structure	A named list. The list names are the factor names. Each element is a character vector of feature names for the corresponding factor. Should be ordered corresponding to the rows of the 'loading' argument.
n_rows	An integer for the number of rows for each independent dataset.
n_datasets	An integer for the number of independent datasets.
seed	An integer for the seed.

Value

Returns a list that contains the complete data and missing data.

Examples

#----------------------------------------------------------------------------
# Data simulation example
#----------------------------------------------------------------------------
# By default, sim() will simulate 3 variables from 3 different domains.
d_sim <- sim()
str(d_sim)
complete_data <- d_sim$complete
missing_data <- d_sim$missing

bklamer/rosetta documentation built on Nov. 4, 2019, 7:31 a.m.