Class of objects that contain information to speed computations
in conditional Gaussian modeling functions that handle missing data.
GENERATION:
This class of objects is returned from the
"preCgm" function.
METHODS:
The class
"preCgm" has methods:
,
,
,
, and
.
STRUCTURE:
The class
"preCgm" object consists of a list containing the
following values:
VALUE:
w
a sorted matrix containing the factors in the model.
n
the number of rows in matrix
w
p
the number of columns in matrix
w
d
vector containing the number of levels in each of the columns in
matrix
w
jmp
vector containing the increment between consecutive cells in the table
for each factor in
w.
z
a sorted matrix contraining the values of the continuous variables in
the model. The number of rows in
z must be
n, the number of rows
in
w.
q
the number of columns in z
r
matrix containing the missing value indicators for the patterns of
missing data in matrix
cbind(w,z). The value
1 indicates a known
value, while
0 indicates a missing value.
rz
matrix containing the missing value indicators for the patterns of
missing data in matrix
z. The value
1 indicates a known value,
while
0 indicates a missing value.
rw
matrix containing the missing value indicators for the patterns of
missing data in matrix
w. The value
1 indicates a known value,
while
0 indicates a missing value.
nmis
number of missing values of each variable in the matrix
c(w,z).
ro
an integer vector such that
cbind(w,z)[ro, ] yields the same order
of data as was observed in the input data.
mdpzgrp
the number of distinct data patterns within a group with the same
missing value pattern for
z.
mdpwgrp
the number of distinct data patterns within a group with the same
missing value pattern for
w.
mdpgst
the number of previous patterns of complete cells prior to this
pattern of missing cells.
mobs
index of the first cell in the subtable for a missing group.
mobsst
index of first observation in a unique pattern.
nmobs
number of observations for each unique pattern.
ncells
total number of cells in the table.
ngrp
the total number of patterns in the data.
npattz
the number of groups of missing data for the discrete variables.
rnames
the row names in the original data.
npsi
the number of distinct covariance matrix parameters.
psi
a matrix containing the locations of the covariance matrix parameters.
xbar
a mean vector subtracted from each observation in computing the
standardized values.
svd
a scale vector. Standardized values are computed for each vector of
observations by subtracting the appropriate means and dividing by the
appropriate scales.
wnames
the names of the variables in
w
znames
the names of the variables in
z.
slevs
a list containig the names associated with each of the levels in each
column of
x.
varnames
the names of the variables in the original data set.