complete {mice} | R Documentation |
Takes an object of type mids, fills in the missing data, and returns the completed data in a specified format.
complete(x, action=1)
x |
An object of class 'mids'
(created by the function mice() ). |
action |
If action is a scalar between 1 and x$m ,
the function returns the data with the action's
imputation filled in. Thus, action=1 returns
the first
completed data set.
The can also be one of the following
strings: "long" , "broad" , "repeated" .
This has the following meaning:
begin{description}
item[action="long" ] produces a long matrix with n*m
rows,containing all imputed data plus two additional
variables "_ID_" (containing the row.names)
and "_IMP_" (containing the imputation number).
item[action="broad" ] produces a broad matrix with m times
the number of columns in the original data.
The first ncol(x$data ) columns contain the first
imputed data matrix. Column names are changed to
reflect the imputation number.
item[action="repeated" ] produces a broad matrix with m times
ncol(x$data) columns. The first m columns
give the filled-in first variable. Column names are
changed to reflect the imputation number.
end{description}
|
A data frame with the imputed values filled in.
Stef van Buuren, Karin Oudshoorn, 2000
Van Buuren, S. & Oudshoorn, C.G.M. (2000). Multivariate Imputation by Chained Equations: MICE V1.0 User's manual. Report PG/VGZ/00.038, TNO Prevention and Health, Leiden.
data(nhanes) imp <- mice(nhanes) # do default multiple imputation on a numeric matrix mat <- complete(imp) # fills in the first imputation mat <- complete(imp, 3) # fills in the third imputation mat <- complete(imp, "long") # produces a long matrix with stacked complete data mat <- complete(imp, "b") # a broad matrix cor(mat) # for numeric mat, produces a blocked correlation matrix, where # each m*m block contains of the same variable pair over different # multiple imputations.