pimp: Permutation Importance

In giuseppec/featureImportance: Model-agnostic feature importance with the mlr package

Description

Implementation of PIMP (see Altmann et al. (2010))

Usage

pimp(
  learner,
  task,
  resampling,
  measures,
  features = NULL,
  n.feat.perm = 1,
  n.target.perm = 1
)

Arguments

learner	[Learner] The learner.
task	(Task) The task.
resampling	(ResampleDesc or ResampleInstance) Resampling strategy. If a description is passed, it is instantiated automatically.
measures	[Measure | list of Measure | function | list of function] Performance measure(s) used to measure the model performance. Can also be a named list of function with signature function(y, pred), where y and pred are vectors containing the true and the predicted values of the target.
features	[list of character] A list where each element contains the names of at least one feature for which the permutation importance should be computed. If a list element contains two or more features, they will be permuted block-wise (without breaking the relationship between those features).
n.feat.perm	[numeric(1)] The number of permutations of the feature(s) used to compute the feature importance. The default is 50. If set to NULL, the cartesian product is used.
n.target.perm	[numeric(1)] The number of permutations of the target used to compute a null distribution for the importance values.

References

Altmann, A., Tolosi, L., Sander, O., & Lengauer, T. (2010). Permutation importance: a corrected feature importance measure. Bioinformatics, 26(10), 1340-1347.

giuseppec/featureImportance documentation built on June 1, 2021, 11:04 a.m.