Marginal Feature Sampler

Implements marginal sampling for PFI, where the feature of interest is sampled independently of other features

Super class

xplainfi::FeatureSampler -> MarginalSampler

Methods

Public methods

• MarginalSampler$new()

• MarginalSampler$sample()

• MarginalSampler$sample_newdata()

• MarginalSampler$clone()

Inherited methods

• xplainfi::FeatureSampler$print()

---

Method new()

Creates a new instance of the MarginalSampler class

Usage

MarginalSampler$new(task)

Arguments

task

(mlr3::Task) Task to sample from

---

Method sample()

Sample from stored task by permutation (marginal distribution).

Usage

MarginalSampler$sample(feature, row_ids = NULL)

Arguments

feature

(character) Feature(s) to sample.

row_ids

(integer() | NULL) Row IDs to use. If NULL, uses all rows.

Returns

Modified copy with permuted feature(s).

---

Method sample_newdata()

Sample from external data by permutation. See $sample() for details.

Usage

MarginalSampler$sample_newdata(feature, newdata)

Arguments

feature

(character) Feature(s) to sample.

newdata

(data.table) External data to use.

Returns

Modified copy with permuted feature(s).

---

Method clone()

The objects of this class are cloneable with this method.

Usage

MarginalSampler$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

library(mlr3)
task = tgen("2dnormals")$generate(n = 100)
sampler = MarginalSampler$new(task)
# Sample using row_ids from stored task
sampled_data = sampler$sample("x1")
# Or use external data
data = task$data()
sampled_data_ext = sampler$sample_newdata("x1", newdata = data)