Compute SHAP-based VI scores for the predictors in a model. See details below.
vi_shap(object, ...)# S3 method for default
vi_shap(object, feature_names = NULL, train = NULL, ...)
A fitted model object (e.g., a "randomForest" object).
Additional optional arguments to be passed on to
explain.
Character string giving the names of the predictor
variables (i.e., features) of interest. If NULL (the default) then the
internal `get_feature_names()` function will be called to try and extract
them automatically. It is good practice to always specify this argument.
A matrix-like R object (e.g., a data frame or matrix)
containing the training data. If NULL (the default) then the
internal `get_training_data()` function will be called to try and extract it
automatically. It is good practice to always specify this argument.
A tidy data frame (i.e., a "tibble" object) with two columns,
Variable and Importance, containing the variable name and its
associated importance score, respectively.
This approach to computing VI scores is based on the mean absolute value of the SHAP values for each feature; see, for example, https://github.com/slundberg/shap and the references therein.
Strumbelj, E., and Kononenko, I. Explaining prediction models and individual predictions with feature contributions. Knowledge and information systems 41.3 (2014): 647-665.