PredictionMultioutput: Prediction Object for MultiOutputer Analysis
Description Super class Active bindings Methods
Description
This object wraps the predictions returned by a learner of class LearnerMultiOutput, i.e. the predicted partition and MultiOutputer probability.
Super class
mlr3::Prediction -> PredictionMultioutput
Active bindings
predictions(
list())
Access the stored predictions.missing(
integer())
Returnsrow_idsfor which the predictions are missing or incomplete.row_ids(
integer())
Access the stored row_ids.
Methods
Public methods
Inherited methods
Method new()
Creates a new instance of this R6 class.
Usage
PredictionMultioutput$new( task = NULL, row_ids = task$row_ids, predictions = list(), check = TRUE, ... )
Arguments
task(TaskMultioutput)
Task, used to extract defaults forrow_ids.row_ids(
integer())
Row ids of the predicted observations, i.e. the row ids of the test set.predictions(
list())
(Named) list of per-target predictions. Used to construct thePrediction-object.check(
logical(1))
IfTRUE, performs argument checks and predict type conversions....(
list())
(Named) list of per-target truths. Only used for compatibility withPrediction$new().
Method print()
Printer for the Prediction object.
Usage
PredictionMultioutput$print(...)
Arguments
...(
any)
Not used.
Method score_separate()
Returns scores for each measure separately.
Usage
PredictionMultioutput$score_separate(measures, task)
Arguments
measureslist
List ofMeasureMultioutputto score.taskTaskMultioutput
Task to use for scoring
Returns
A numeric() vector of scores.
Method clone()
The objects of this class are cloneable with this method.
Usage
PredictionMultioutput$clone(deep = FALSE)
Arguments
deepWhether to make a deep clone.
