Recompute QINI Curves Using IPW Scores

Recomputes QINI curves for binary treatment causal forest models using Inverse Probability Weighted (IPW) scores. This is a simpler alternative to margot_recompute_qini_aipw() that uses IPW scores only.

Usage

margot_recompute_qini_ipw(
  margot_result,
  model_name = NULL,
  treatment_var = NULL,
  W.hat = NULL,
  verbose = TRUE
)

Arguments

margot_result

A list returned by margot_causal_forest().

model_name

Character string specifying which model to recompute. If NULL (default), all models will be recomputed.

treatment_var

Character string specifying the treatment variable name. If NULL, the function will try to detect it automatically.

W.hat

Numeric vector of propensity scores. If NULL, will use propensity scores from the model or estimate them.

verbose

Logical. If TRUE, prints progress messages. Default is TRUE.

Value

A modified version of the input margot_result with updated QINI curves based on IPW scores. The structure remains compatible with all existing plotting and interpretation functions.

Details

This function provides a way to recompute QINI curves using the modern maq API with IPW scores. It's useful for:

• Debugging QINI curve differences

• Faster computation compared to AIPW

• Ensuring consistency across different model computations

The function handles various data storage patterns including models stored in the results list or in a separate full_models list.

Examples

if (FALSE) { # \dontrun{
# Recompute QINI curves with IPW for all models
results_ipw <- margot_recompute_qini_ipw(margot_results)

# Recompute for a specific model
results_ipw <- margot_recompute_qini_ipw(
  margot_results,
  model_name = "model_anxiety"
)

# Specify treatment variable if auto-detection fails
results_ipw <- margot_recompute_qini_ipw(
  margot_results,
  treatment_var = "t1_treatment"
)

# Compare with AIPW results
results_aipw <- margot_recompute_qini_aipw(margot_results)
# Plot both for comparison
} # }