complete_results: Completed data sets as long data.frame

In stephematician/miForang: Single or multiple imputation of missing data using random forests

Description

Completed (imputed) data sets given incomplete data passed to smirf and the output thereof, returned in a long-format data.frame.

Usage

complete_results(X, mf, k = length(mf$results),
  to.complete = names(mf$which_imputed), replace = F)

Arguments

X	data.frame; incomplete data set passed to smirf in construction of mf.
mf	list; imputed data sets and output of missForest procedure as returned by smirf.
k	integer; number of completed data sets to generate, default is the number of imputed data sets in mf.
to.complete	character; vector of names of variables to include, default is all variables in imputed data sets in mf.
replace	boolean; indicator of sampling from imputed data sets with replacement or not.

Details

Description

Value

data.frame; completed long-format data containing the columns of X and two more columns,

.source

integer; id of the source data set in mf of the imputed data in the row (if any), otherwise NA, and;

.imp

integer; id of the complete data set between 1 and k, or NA if it belongs to all (i.e. contains no imputed data),

and each row is either

• a complete case whose values are unaltered, or;

• an incomplete case with a sample from the imputed data replacing any missing values.

To-do

• document complete_results

• write tests for complete_results

See Also

smirf

stephematician/miForang documentation built on July 23, 2019, 5:11 p.m.