Splitting methods — make_unconditional

These functions handle split routines used in the estimation of crosswalks. In general, we recommend using these functions via crosswalk(), but they are made available to the user who wishes to call them directly.

make_unconditional_splits() : implements the unconditional split routine
make_conditional_splits() : implements the conditional split routine, in which splits in the data are defined by a binary variable (e.g., dementia) correlated with given cognitive measures (e.g., MMSE and MoCA) because the underlying cognitive construct is a common cause of all three
make_splits() : a handler that checks the input dataset and passes it on to the appropriate split function with relevant arguments specified

Usage

make_unconditional_splits(data, niter = NULL)

make_conditional_splits(cdvar = NULL, data, loop = FALSE)

make_splits(cdvar = NULL, data, cdloop = FALSE, niter = NULL)

Arguments

data: A data.table, data.frame, matrix, or list containing the cognitive measure data
niter: Number of iterations to conduct for an unconditional split routine
cdvar: Character string naming auxiliary variable by which to condition splits
loop: Boolean declaring whether to use a for loop. The default FALSE will generate splits and operate on an expanded data.table in-memory. If your machine has limited memory, set this argument to TRUE in order to process splits sequentially. The default option should be much faster.
cdloop: Boolean passed to loop argument of make_conditional_splits(). Ignored when conducting unconditional splits.