extract_pd_nestedrf: Extract partial dependence (and performance) from a trained...

Description Usage Arguments Details Value Warning documentation to-do See Also

Description

Computes the marginal relationship between a subset of the predictors (here, two variables at a time) and the model’s predictions by averaging over the marginal distribution of the compliment of this subset of the predictors, taking in account the interaction between the chosen predictors.

Usage

1
2
3
4
5
6
7
8
extract_pd_nestedrf(
  learner_id = 1,
  in_rftuned,
  datdf,
  selcols,
  nvariate,
  ngrid
)

Arguments

learner_id

(integer) Index of the outer resampling instance to be analyzed.

in_rftuned

ResampleResult from a classification RF.

datdf

Data from the task that was used to train RF.

selcols

Character vector of the predictor variables to analyze.

ngrid

(integer) Number of values of the predictor variables over which to compute the marginal relationship.

Details

Also accept learners of class GraphLearner.
Uses partial_dependence for computing.

Value

A data.table with the following columns.

value1
value2
0

is perennial) at value1 and value2

1

is intermittent) at value1 and value2

[perf]
var1
var2

Warning

Has only been tested on mlr_learners_classif.ranger

documentation to-do

Can add an example down the line, add source.

See Also

weighted_vimportance_nestedrf, ggpd_bivariate


messamat/globalIRmap documentation built on July 4, 2021, 10:48 a.m.