fs_betaHhat_table: Deduplicated 'beta(Hhat)' targets over distinct realized...

View source: R/betaHhat_truth.R

fs_betaHhat_tableR Documentation

Deduplicated beta(Hhat) targets over distinct realized rules

Description

Scores each distinct realized rule once. Deduplication is by rule, not by replicate: two replicates landing on the same rule score the same target, and the target does not depend on the trial.

Usage

fs_betaHhat_table(
  sg_defs,
  frame,
  focus,
  outcome_type = c("survival", "binary", "continuous", "count"),
  effect_measure = NULL,
  outcome.name = "y_sim",
  event.name = "event_sim",
  treat.name = "treat_sim",
  sg_def_structs = NULL,
  method = c("frame", "exact"),
  dgm = NULL,
  analysis_time = 84,
  max_entry = 24,
  cens_adjust = log(1.5)
)

Arguments

sg_defs

Character vector of realized rules, one entry per replicate. NA or "" marks a replicate where no subgroup was identified.

frame

Data frame. The fixed evaluation population.

focus, outcome_type, effect_measure, outcome.name, event.name, treat.name

Passed to fs_betaHhat_one().

sg_def_structs

Optional named list of structured GRF definitions, keyed by the rule string, used in preference to string parsing.

method, dgm, analysis_time, max_entry, cens_adjust

Passed to fs_betaHhat_one(); under method = "exact" the analytic population is built once for all rules.

Value

A data frame keyed by sg_def, one row per distinct rule, with the fs_betaHhat_one() schema. Resolution counters are attached as attributes and retrievable with fs_betaHhat_counts().

Resolution accounting

Six counters are attached to the result and are the fix for targets being dropped silently. n_eff reported beside a target is what makes a reduced denominator visible; a coverage figure computed on an unknown fraction of replicates is not interpretable, and the fraction has to travel with it.

A seventh, n_reps_undetected, counts replicates with no rule at all, so that n_reps_resolved + n_reps_unresolved + n_reps_undetected closes against n_reps_total.


forestsearch documentation built on Nov. 15, 2026, 9:06 a.m.