learn_threshold: Learn Optimal Decision Threshold

In PLUCR: Policy Learning Under Constraint

Learn Optimal Decision Threshold

Description

This function estimates the optimal decision threshold for treatment assignment using cross-fitted nuisance function estimates and targeted maximum likelihood estimation (TMLE). The procedure evaluates candidate thresholds based on empirical performance criteria, selecting the threshold that maximizes the policy value under constraint satisfaction.

Usage

learn_threshold(theta, X, A, Y, Xi, folds, alpha)

Arguments

theta	A numeric matrix (k x d). Each row is from FW inner minimization, used to recover an extremal point for convex function construction.
X	A matrix or data frame of covariates of size n x d (input data in ⁠[0,1]⁠).
A	A binary vector or matrix of length n indicating treatment assignment (0 or 1).
Y	A numeric vector or matrix of length n representing primary outcomes (in ⁠[0,1]⁠).
Xi	A numeric vector or matrix of adverse events outcomes.
folds	A list of cross-validation folds (e.g., a list of indices for each fold).
alpha	A numeric scalar representing the constraint tolerance (in ⁠[0,1/2]⁠, 0.1 by default).

Value

A numeric value corresponding to the optimal threshold chosen from the candidate sequence.

PLUCR documentation built on March 30, 2026, 5:08 p.m.