Compare imputed values to ground truth at originally missing positions,
by calling the Python function evaluate_imputation() defined in the
ciss_vae module.
Usage
evaluate_imputation(imputed_df, df_complete, df_missing)
Arguments
- imputed_df
A data.frame or tibble of imputed values (same dim as df_complete)
- df_complete
A data.frame or tibble of the complete (ground-truth) values
- df_missing
A data.frame or tibble with NAs indicating the original missing entries
Value
A list with components:
mseMean squared error at the originally missing positions
comparisonA data.frame with columns row, col,
true, imputed, and squared_error