Report Policy Value for Consensus Trees — margot_report_consensus_policy

Computes and reports how much better the consensus policy trees perform than (a) treat-all (ATE baseline) and (b) treat-none (universal control), with 95

Usage

margot_report_consensus_policy_value(
  object,
  model_names = NULL,
  depths = c(1L, 2L),
  R = 499L,
  seed = 42L,
  ci_level = 0.95,
  include_treated_only = FALSE,
  label_mapping = NULL,
  verbose = TRUE
)

Arguments

object: Object of class "margot_stability_policy_tree" produced by [margot_policy_tree_stability()]. Must contain consensus policy trees in `policy_tree_depth_1` and/or `policy_tree_depth_2`, plus `plot_data` and `dr_scores`/`dr_scores_flipped`.
model_names: Optional character vector of model names to include (with or without `model_` prefix). Default: all.
depths: Integer vector of depths to evaluate (default `c(1, 2)`).
R: Integer ≥ 199; number of bootstrap replicates (default 499).
seed: Integer or NULL; RNG seed for reproducibility (default 42).
ci_level: Confidence level for intervals (default 0.95).
verbose: Logical; print progress (default TRUE).

Value

A data.frame with one row per model × depth × contrast containing estimate, standard error, and confidence interval.