Class of objects that contain information to speed computations
in multivariate normal modeling functions that handle missing data.
GENERATION:
This class of objects is returned from the
"preGauss" function.
METHODS:
The class
"preGauss" has methods:
,
,
,
,
, and
.
STRUCTURE:
The class
"preGauss" object consists of a list containing the
following values:
VALUE:
table
a matrix containing the number of complete and missing rows and columns.
partial.var
the cholesky decomposition of the columns of complete data, augmented
with a column of ones for computing the means. Also included is a
vector of residuals from the Given's rotations used in computing the
Cholesky decomponsition.
incomplete.rows
a matrix containing the (sorted) rows of data with missing values.
column.info
a list containing the sums, means, and variances of the columns of
data. Also included is the original column index of the variables.
missing.pattern
a list containing a matrix of missing value patterns, the number of
times each row of patterns is replicated, and the index of the first,
and last, missing value in the row.
varnames
vector of variable names as specified in the original data.
value.
complete.rows
matrix containing the rows of data with no missing values.