| imp.cat {cat} | R Documentation |
Performs single random imputation of missing values in a categorical dataset under a user-supplied value of the underlying cell probabilities.
imp.cat(s, theta)
s |
summary list of an incomplete categorical dataset created by the
function prelim.cat.
|
theta |
parameter value under which the missing data are to be imputed.
This is an array of cell probabilities of dimension s$d whose
elements sum to one, such as produced by em.cat, ecm.cat,
da.cat, mda.cat or dabipf.
|
Missing data are drawn independently for each observational unit from
their conditional predictive distribution given the observed data and
theta.
If the original incomplete dataset was in ungrouped format
(s$grouped=F), then a matrix like s$x except that all NAs have
been filled in.
If the original dataset was grouped, then a list with the following
components:
x |
Matrix of levels for categorical variables
|
counts |
vector of length nrow(x) containing frequencies or counts
corresponding to the levels in x.
|
IMPORTANT: The random number generator seed must be set by the
function rngseed at least once in the current session before this
function can be used.
prelim.cat, rngseed, em.cat, da.cat, mda.cat, ecm.cat,
dabipf
data(crimes) x <- crimes[,-3] counts <- crimes[,3] s <- prelim.cat(x,counts) # preliminary manipulations thetahat <- em.cat(s) # find ML estimate under saturated model rngseed(7817) # set random number generator seed theta <- da.cat(s,thetahat,50) # take 50 steps from MLE ximp <- imp.cat(s,theta) # impute once under theta theta <- da.cat(s,theta,50) # take another 50 steps ximp <- imp.cat(s,theta) # impute again under new theta