create_marginal_data_training {shapr} | R Documentation |
Function that samples data from the empirical marginal training distribution
Description
Sample observations from the empirical distribution P(X) using the training dataset.
Usage
create_marginal_data_training(
x_train,
n_explain,
Sbar_features,
n_MC_samples = 1000,
stable_version = TRUE
)
Arguments
x_train |
Data.table with training data. |
Sbar_features |
Vector of integers containing the features indices to generate marginal observations for.
That is, if |
stable_version |
Logical. If |
Value
Data table of dimension n_MC_samples
\times
length(Sbar_features)
with the sampled observations.
Author(s)
Lars Henry Berge Olsen