mlr3classifiercv

cross-validate a mlr3 classification model

mlr3classifiercv(
  dfin,
  tcols,
  nrepeats = 10,
  partrate = 0.8,
  dup_size = 0,
  balancing = "smote",
  mylearners = mlr3classifiers(),
  subjectIDs = NULL,
  verbose = TRUE
)

Arguments

dfin

dataframe input

tcols

columns for the prediction task - first is the target outcome

nrepeats

number of subsampling-driven train test runs

partrate

partition ratio for the training 0.8 equals 80 percent train 20 test

dup_size

integer for over/under/smote sampling

balancing

string over, under, smote, none are the options

mylearners

the mlr3 learners over which to search ; defaults to those that support multiclass

subjectIDs

unique IDs per subject; aids with repeated measurement data by ensuring subjects exist in uniquely in either train or test

verbose

boolean

Value

dataframe quantifying performance

Author

Avants BB