Fit a CRR Model — fit

High-level wrapper for fitting conditional risk-ratio models with non-negative outcomes. The wrapper supports both single-method fits and cross-validated portfolios across multiple learner families.

Usage

fit_crr(
  task,
  method = supported_crr_methods(),
  base_learner = get_autoML(),
  treatment_level = NULL,
  control_level = NULL,
  cross_validate = NULL,
  cv_control = NULL,
  sieve_num_basis = NULL,
  sieve_basis_grid = NULL,
  sieve_interaction_order = 3
)

Arguments

task: An hte3_Task.
method: Either a single method or a character vector of methods drawn from "ep", "ipw", and "t". A vector creates a learner portfolio that is selected by cross-validation.
base_learner: Base supervised learner used by the meta-learner.
treatment_level: Optional treated level used for the contrast.
control_level: Optional control level used for the contrast.
cross_validate: Whether to cross-validate across candidate learners. Defaults to TRUE for stacks and is automatically enabled for learner portfolios.
cv_control: Optional sl3 cross-validation control list.
sieve_num_basis: Optional sieve basis size for EP-learner fits.
sieve_basis_grid: Optional EP basis-size grid used when method includes "ep" and wrapper-level cross-validation is enabled. If NULL, the wrapper uses the heuristic default grid c(d, 2d, 4d, 6d, 8d), where d is the number of modifiers.
sieve_interaction_order: Interaction order for EP-learner sieve construction.

Value

An object of class hte3_model.

Details

CRR models are supported for binary or categorical treatment tasks and require a non-negative outcome.

For EP-learner fits, sieve_num_basis controls a single fixed sieve size when cross_validate = FALSE. When cross_validate = TRUE, sieve_basis_grid controls which EP basis sizes are compared by the wrapper. After fitting, summary() reports the selected candidate and the selected EP basis size when applicable.

Examples

if (FALSE) { # \dontrun{
library(sl3)

data <- hte3_example_data(n = 80, seed = 1)
task <- hte_task(
  data = data,
  modifiers = c("W1", "W2"),
  confounders = c("W1", "W2", "W3"),
  treatment = "A",
  outcome = "Y_binary",
  propensity_learner = Lrnr_mean$new(),
  outcome_learner = Lrnr_mean$new(),
  mean_learner = Lrnr_mean$new(),
  cross_fit = FALSE
)

fit <- fit_crr(
  task,
  method = "ep",
  base_learner = Lrnr_mean$new(),
  cross_validate = TRUE,
  cv_control = list(V = 2),
  sieve_basis_grid = c(3, 6, 9)
)

summary(fit)
} # }