Skip to contents

Compute variable importance using SHAP values

Source: R/machine_learning.R
compute_variable.importance.Rd

Computes the variable importance for a machine learning model using SHAP (SHapley Additive exPlanations) values.

Usage

compute_variable.importance(model, stacking = FALSE, n_cores = 2)

Arguments

model

A trained machine learning model.

stacking

A logical value indicating whether the model was trained using stacking (default is FALSE).

n_cores

An integer specifying the number of workers to use for parallel computation (default is 2).

Value

A data frame with SHAP values representing the variable importance for each feature.

Details

If stacking is TRUE, the function computes the SHAP values for each base model in the stacked ensemble model and averages them. If stacking is FALSE, the function computes the SHAP values for the provided single machine learning model. The computed SHAP values are returned as a data frame with features as rows and samples as columns.