Apply a (nested list) of rROC results to new data

Restriction returns an informative range which tells that some samples should be discarded according to their predictor value. This function applies the restriction to a new dataset. Predictor values outside the informative range are imputed with removed_impute.

Usage

apply_restriction(object, newdata, feature = NA, removed_impute = -1)

Arguments

object

A (nested list) of rROC results. See rROC for details.

newdata

A data frame with all predictor variables.

feature

The current feature. If object is a single restrictedROC object, this is the name of the feature. If object is a (nested list), this is the name of the current list element. Example: list("outcome_1"=list("feature_1"=rroc_result)) feature is "feature_1" when rroc_result is processed when rroc_result is detected.

removed_impute

The value to impute for samples outside the informative range.

Value

A list with three elements:

• predictions: A data frame with the original predictor values and the imputed values. It has three columns:

  predictor

  is the original predictor value.

  keep

  is a logical vector that indicates whether the sample is inside the informative range.

  bounded

  is the imputed value.

• thresholds: A data frame with the feature, threshold, the part of the predictor that was used and the imputed value for predictor values outside the informative range. The informative range is then:

  low

  (-Inf, threshold)

  high

  (threshold, Inf)