| cbind.mids {mice} | R Documentation |
mids object.This function combines two mids objects columnwise into a single
object of class mids, or combines a mids object with a vector,
matrix, factor or data.frame columnwise into an object of class mids.
The number of rows in the (incomplete) data x$data and y (or
y$data if y is a mids object) should be equal. If
y is a mids object then the number of imputations in x
and y should be equal. Note: If y is a vector or factor its
original name is lost and it will be denoted with y in the mids
object.
cbind.mids(x, y, ...)
x |
A |
y |
A |
... |
Additional |
An S3 object of class mids
Component call is a vector, with first argument the mice() statement
that created x and second argument the call to cbind.mids().
Component data is the codecbind of the (incomplete) data in x$data
and y$data. Component m is the number of imputations.
Component nmis is an array containing the number of missing observations per
column.
Component imp is a list of nvar components with the generated multiple
imputations. Each part of the list is a nmis[j] by m matrix of
imputed values for variable j. The original data of y will be
copied into this list, including the missing values of y then y
is not imputed.
Component method is a vector of strings of length(nvar) specifying the
elementary imputation method per column. If y is a mids object this
vector is a combination of x$method and y$method, otherwise
this vector is x$method and for the columns of y the method is
set to ''.
Component predictorMatrix is a square matrix of size ncol(data)
containing integer data specifying the predictor set. If x and
y are mids objects then the predictor matrices of x and
y are combined with zero matrices on the off-diagonal blocks.
Otherwise the variables in y are included in the predictor matrix of
x such that y is not used as predictor(s) and not imputed as
well.
Component visitSequence is the sequence in which columns are visited. The same
as x$visitSequence.
Component seed is the seed value of the solution, x$seed.
Component iteration is the last Gibbs sampling iteration number,
x$iteration.
Component lastSeedValue is the most recent seed value, x$lastSeedValue
Component chainMean is the combination of x$chainMean and
y$chainMean. If y$chainMean does not exist this element equals
x$chainMean.
Component chainVar is the combination of x$chainVar and y$chainVar.
If y$chainVar does not exist this element equals x$chainVar.
Component pad is a list containing various settings of the padded imputation
model, i.e. the imputation model after creating dummy variables. This list
is defined by combining x$pad and y$pad if y is a
mids object. Otherwise, it is defined by the settings of x and
the combination of the data x$data and y.
Component loggedEvents is set to x$loggedEvents.
If a column of y is categorical this is ignored in the
padded model since that column is not used as predictor for another column.
Karin Groothuis-Oudshoorn, Stef van Buuren, 2009
# append 'forgotten' variable bmi to imp temp <- boys[,c(1:3,5:9)] imp <- mice(temp,maxit=1,m=2) imp2 <- cbind.mids(imp, data.frame(bmi=boys$bmi)) # append maturation score to imp (numerical) mat <- (as.integer(temp$gen) + as.integer(temp$phb) + as.integer(cut(temp$tv,breaks=c(0,3,6,10,15,20,25)))) imp2 <- cbind.mids(imp, as.data.frame(mat)) # append maturation score to imp (factor) # known issue: new column name is 'y', not 'mat' mat <- as.factor(mat) imp2 <- cbind.mids(imp, mat) # append data frame with two columns to imp temp2 <- data.frame(bmi=boys$bmi,mat=as.factor(mat)) imp2 <- cbind.mids(imp, temp2) # combine two mids objects impa <- mice(temp, maxit=1, m=2) impb <- mice(temp2, maxit=2, m=2) # first a then b impab <- cbind.mids(impa, impb) # first b then a impba <- cbind.mids(impb, impa)