Corrected-T CI

Corrected-T confidence intervals based on ResamplingSubsampling. A heuristic factor is applied to correct for the dependence between the iterations. The confidence intervals tend to be liberal. This inference method can also be applied to non-decomposable losses.

Parameters

Only those from MeasureAbstractCi.

References

Nadeau, Claude, Bengio, Yoshua (1999). “Inference for the generalization error.” Advances in neural information processing systems, 12.

Super classes

mlr3::Measure -> mlr3inferr::MeasureAbstractCi -> MeasureCiCorrectedT

Methods

Public methods

• MeasureCiCorrectedT$new()

• MeasureCiCorrectedT$clone()

Inherited methods

• mlr3::Measure$format()
• mlr3::Measure$help()
• mlr3::Measure$print()
• mlr3::Measure$score()
• mlr3inferr::MeasureAbstractCi$aggregate()

---

Method new()

Creates a new instance of this R6 class.

Usage

MeasureCiCorrectedT$new(measure)

Arguments

measure

(Measure or character(1))
A measure of ID of a measure.

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MeasureCiCorrectedT$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

m_cort = msr("ci.cor_t", "classif.acc")
m_cort
#> <MeasureCiCorrectedT:classif.acc>: Corrected-T CI
#> * Packages: mlr3, mlr3measures, mlr3inferr
#> * Range: [0, 1]
#> * Minimize: FALSE
#> * Average: custom
#> * Parameters: alpha=0.05, within_range=TRUE
#> * Properties: primary_iters
#> * Predict type: response
rr = resample(
  tsk("sonar"),
  lrn("classif.featureless"),
  rsmp("subsampling", repeats = 10)
)
rr$aggregate(m_cort)
#>       classif.acc classif.acc.lower classif.acc.upper 
#>         0.4898551         0.3408775         0.6388327