Estimate the influence function for the given measure of predictiveness.
cv_predictiveness_update(
fitted_values,
y,
folds,
weights = rep(1, length(y)),
type = "r_squared",
na.rm = FALSE
)fitted values from a regression function; a list of length V, where each object is a set of predictions on the validation data.
the outcome.
the cross-validation folds
weights for the computed influence curve (e.g., inverse probability weights for coarsened-at-random settings)
which risk parameter are you estimating (defaults to r_squared, for the $R^2$)?
logical; should NAs be removed in computation? (defaults to FALSE)
The estimated influence function values for the given measure of predictiveness.
See the paper by Williamson, Gilbert, Simon, and Carone for more details on the mathematics behind this function and the definition of the parameter of interest.