Add predictions from an AutoML object to a test set

This function generates predictions from an AutoML object for the test set in a train-test split, and adds the predictions to the test set. Confidence levels can also optionally be added alongside predictions.

Usage

add_predictions_from_automl(automl, tt, tt_h2o, add_confidence_level = TRUE)

Arguments

automl

H2O AutoML object (from h2o::h2o.automl() or get_automl_model())

tt

train-test list (see ttsplit())

tt_h2o

H2O pointers to train and test frames (see ship_train_and_test_to_h2o())

add_confidence_level

(logical) whether to add confidence levels alongside predictions in test set (default TRUE)

Value

train-test list with predictions (and optionally confidence levels) added to $test

See also

h2o::h2o.predict(), h2o::h2o.automl(), get_automl_model()

Examples

if (FALSE) {
init_h2o()
tt <- df |> ttsplit()
tt_h2o <- tt |> ship_train_and_test_to_h2o()
automl <- tt_h2o |> get_automl_model(target="y1")
tt <- add_predictions_from_automl(automl)
}