Module policy 

Policy trees: a shallow decision tree that recommends treatment where the estimated effect is positive, as DoubleML’s DoubleMLPolicyTree (Athey and Wager, 2021).

The tree classifies the sign of the orthogonal signal ψ_b (labels (sign + 1) / 2) with sample weights |ψ_b|, so misclassifying units with large effects costs more. It is grown like scikit-learn’s DecisionTreeClassifier (Gini criterion, depth and leaf-size limits) and then pruned by minimal cost-complexity pruning (ccp_alpha), with the same pruning order as scikit-learn.

Structs

PolicyTree

A fitted policy tree.

Enums

PolicyNode

One node of a fitted PolicyTree (see PolicyTree::nodes).