interpret_search_config: Interpret and Diagnose Search Configuration

View source: R/interpret_search_config.R

interpret_search_configR Documentation

Interpret and Diagnose Search Configuration

Description

Generates a human-readable diagnostic of how ForestSearch will interpret the user's parameter combination. Prints via message() and returns an invisible list of diagnostics.

Usage

interpret_search_config(
  outcome_type,
  effect_measure,
  adverse_outcome,
  effect_threshold,
  consistency_threshold,
  use_lasso,
  use_grf,
  outcome.name,
  event.name,
  treat.name,
  offset.name = NULL,
  subgroup_method = "consistency",
  quiet = FALSE
)

Arguments

outcome_type

Character. One of "survival", "binary", "continuous", "count".

effect_measure

Character. "HR", "OR", "RR", "RD", "MD", "IRR", "IRD".

adverse_outcome

Logical. As resolved by forestsearch().

effect_threshold

Numeric. Resolved screening threshold (on log scale for ratio measures, identity for others).

consistency_threshold

Numeric. Resolved consistency threshold.

use_lasso

Logical.

use_grf

Logical. Whether GRF candidate-cut generation is on; see forestsearch.

outcome.name

Character. Name of outcome column.

event.name

Character. Name of event column.

treat.name

Character. Name of treatment column.

offset.name

Character or NULL.

subgroup_method

Character, the resolved subgroup_method. "dina" and "grf" return before the consistency stage and never read the consistency threshold, so the echoed value is annotated as unused on those paths. Display only; nothing is computed from it. Default "consistency".

quiet

Logical. If TRUE, suppress output.

Value

Invisible list with diagnostic fields.


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