PPI estimator for discrete predictions

Estimates the target mean by debiasing the test-set predictions using the calibration-set mean correction. $$\hat\theta = \overline{\hat Y}_{\mathrm{test}} + \left(\overline{Y}_{\mathrm{cal}} - \overline{\hat Y}_{\mathrm{cal}}\right).$$

Usage

ppi_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.