Assemble the Final Imputed Data From the Winning Methods
Source:R/imputation_rse.R
build_best_imputed_data.RdBuilds one data frame in which every variable comes from the imputation method that won for it ([tbl_best_imputation_method()]).
Usage
build_best_imputed_data(
data,
imputations,
best,
require_beats_null = FALSE,
fallback = NULL
)Arguments
- data
The original data frame (with its missing values). Columns that are not in `best` are returned as they are.
- imputations
A **named list** of data frames, one per method, each the same size as `data` and in the same row order (this is not checked beyond the number of rows) with the columns imputed by that method. The names are the method names used in `best`.
- best
A [tbl_best_imputation_method()] result.
- require_beats_null
If `TRUE`, a variable whose best method does not beat the null baseline (RSE of 1 or more) is left with the `fallback` method instead of the winner.
- fallback
Method name (in `imputations`) used for variables without a winner and, with `require_beats_null`, for those that do not beat the null baseline. `NULL` (default) leaves such variables as in `data`.