MeasureMultioutputWeightedAvg: Multioutput weighted average Measure

In mlr-org/mlr3multioutput: Multioutput extension for 'mlr3'

Description

Computes a weighted average over measured scores for each target.

• measures that should be maximized are automatically multiplied by '-1' internally during aggregation and minimize is therefore set to TRUE

• task_type is set to "multioutput".

• Possible values for predict_type are all values from mlr_reflections$learner_predict_types. They are currently collected by accessing each Measures "predict_type" slot. Currently limited to only a single 'predict_type' across all measures.

• packages are all packages required from the supplied measures.

• The range is automatically computed based on the specified weights.

Predefined measures can be found in the mlr3misc::Dictionary mlr3::mlr_measures.

Super classes

mlr3::Measure -> mlr3multioutput::MeasureMultioutput -> MeasureMultioutputWeightedAvg

Public fields

measures

(list())
Access the stored measures.

weights

(numeric())
Access the stored weights.

Methods

Public methods

• MeasureMultioutputWeightedAvg$new()

• MeasureMultioutputWeightedAvg$score_separate()

• MeasureMultioutputWeightedAvg$clone()

Inherited methods

• mlr3::Measure$aggregate()

• mlr3::Measure$format()

• mlr3::Measure$help()

• mlr3::Measure$print()

• mlr3::Measure$score()

---

Method new()

Creates a new instance of this R6 class.

Usage

MeasureMultioutputWeightedAvg$new(
  name = "weightedavg",
  measures = get_default_measures(),
  weights = NULL
)

Arguments

name

(character(1))
Name of the measure. Default: "weightedavg".

measures

(list)
Named list of measures to be applied to each "target".
Either named with target_names, mapping targets to measures or named with task_types, defining one measure per task_type. Defaults to mlr_reflections$default_measures for each task type.

weights

(numeric())
Named numeric defining weights for respective 'measures'.
Either named with target_names, mapping targets to weights or named with task_types, defining one weight per task_type. Defaults to 1 for each measure.

Returns

A MeasureMultioutput

---

Method score_separate()

Returns scores for each measure in self$measures separately.

Usage

MeasureMultioutputWeightedAvg$score_separate(prediction, task, ...)

Arguments

prediction

PredictionMultioutput
Prediction to score.

task

TaskMultioutput
Task to score.

...

(any)
Currently not used.

Returns

A numeric() vector of scores.

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MeasureMultioutputWeightedAvg$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

See Also

Example measures:

mlr-org/mlr3multioutput documentation built on Nov. 22, 2020, 1:17 p.m.