extract_impperf_nestedrf: Extract variable importance and performance from a trained RF...

In NaiaraLopezRojo/globalIRmap: What the Package Does (One Line, Title Case)

Description

Computes the performance, variable importance and associated p-value from either a trained AutoTuner or a trained Learner.

Usage

extract_impperf_nestedrf(
  in_rflearner,
  in_task,
  imp = TRUE,
  perf = TRUE,
  pvalue = TRUE,
  pvalue_permutn = 100
)

Arguments

in_rflearner	An AutoTuner or a trained Learner.
imp	(logical) Whether to compute variable importance.
perf	(logical) Whether to compute the performance measure.
pvalue	(logical) Whether to compute p-values.
pvalue_permutn	(integer) number of permutations to use in p-value calculations

Details

Accepts both a trained AutoTuner or a trained \ link[mlr3]Learner of class mlr_learners_classif.ranger or mlr_learners_classif.cforest.

Also accept learners of class GraphLearner

If p-value==TRUE and mlr_learners_classif.ranger, then compute p-values with importance_pvalues with Altmann permutation method using pvalue_permutn permutations.

If mlr_learners_classif.cforest is provided, p-value is ignored for now.

Value

A data.table with, when all logicals are TRUE, the following columns.

varnames

imp

pvalue

[perf]

documentation to-do

Can add an example down the line, add source.

See Also

weighted_vimportance_nestedrf

NaiaraLopezRojo/globalIRmap documentation built on Dec. 17, 2021, 5:19 a.m.