Converts a trained mlr3 learner into a DALEX explainer for interpretability.
Usage
xai_explainer_mlr3(learner, task, data = NULL, label = NULL, features = NULL)
Arguments
- learner
A trained mlr3 Learner (must have predict_type = "prob")
- task
The mlr3 task used for training
- data
Optional: explicit data for explainer (if NULL, uses task data)
- label
Label for the explainer (default: learner ID)
- features
Optional: subset of features to include (speeds up explanations)
Value
A DALEX explainer object
Examples
if (FALSE) { # \dontrun{
if (requireNamespace("DALEX", quietly = TRUE) &&
requireNamespace("mlr3learners", quietly = TRUE) &&
requireNamespace("ranger", quietly = TRUE)) {
learner <- lrn("classif.ranger", predict_type = "prob")
learner$train(task)
explainer <- xai_explainer_mlr3(learner, task)
}
} # }