pmm_observed_scores: Predicted donor scores for true PMM

In VIM: Visualization and Imputation of Missing Values

Predicted donor scores for true PMM

Description

Scores the observed rows of a target with the trained learner so that PMM donors are matched on their predicted values (Little 1988), not on their observed values. For a ranger fit without bootstrap resampling the honest out-of-bag predictions are used; otherwise the trained learner predicts the observed rows (aregImpute-style when the fit is a bootstrap refit). Returns NULL when the rows cannot be scored.

Usage

pmm_observed_scores(
  learner,
  data_temp,
  obs_idx,
  feature_cols,
  target_col,
  factor_levels,
  method_var,
  boot,
  lhs_transformation = NULL
)

VIM documentation built on Sept. 2, 2026, 5:07 p.m.