Validate the ALASCA model LMM — do

This function performs leave-one-out robustness testing of your ALASCA model. If you didn't specify the number of runs n_validation_runs when initializing the model (see ALASCA), you can do it by running for example model$n_validation_runs <- 100 prior to calling validate. Your dataset is divided into n_validation_folds partitions, keeping group proportions, and one of these are left out. n_validation_folds is set the same way as n_validation_runs.

do_validate()

Arguments

object: An ALASCA object
participant_column: The name of the column containing participant identifier. Needed if not set during initialization of the model.
validate_regression: Whether to validate regression models

Value

An ALASCA object

Examples

load("PE.Rdata")
#> Warning: cannot open compressed file 'PE.Rdata', probable reason 'No such file or directory'
#> Error in readChar(con, 5L, useBytes = TRUE): cannot open the connection
model$n_validation_runs <- 10
#> Error: object 'model' not found
model.val <- validate(model, participant_column = "ID")
#> Error in validate(model, participant_column = "ID"): could not find function "validate"