EIF estimator with GAM calibration

Fits \(g(\hat{Y})\) using a GAM with smooth term. For discrete \(\hat{Y}\) with few categories, this behaves similarly to per-category.

Usage

eif_gam_discrete_point_and_ci(Y_cal, Yhat_cal, Yhat_test, alpha = 0.1)

Arguments

Y_cal

Vector of true labels on the calibration set.

Yhat_cal

Vector of surrogate predictions on the calibration set.

Yhat_test

Vector of surrogate predictions on the test set.

alpha

Miscoverage level for the confidence interval.