Class "preLoglin"
DESCRIPTION:
Classes of objects that contain information to speed computations
in log-linear modeling functions that handle missing data.
GENERATION:
This class of objects is returned from the
"preLoglin"
function.
METHODS:
The class
"preLoglin"
has methods:
,
,
,
, and
.
STRUCTURE:
The class
"preLoglin"
object consists of a list containing the
following values:
VALUE:
- x
-
a sorted matrix containing the factors in the model.
- n
-
the number of rows in matrix
x
- p
-
the number of columns in matrix
x
- d
-
vector containing the number of levels in each of the columns in
matrix
x
- jmp
-
vector containing the increment between consecutive cells in the table
for each factor in
x
.
- r
-
matrix containing the missing value indicator for the matrix
x
. The
value
1
indicates a known value, while
0
indicates a missing
value.
- nmis
-
number of missing values of each variable in the matrix
x
.
- ro
-
an integer vector such that x[ro, ] yields the same order of data as
was observed in the input data.
- mdpgrp
-
the number of distinct patterns cells within a group with the same
missing value pattern.
- 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 observation for each unique pattern.
- ncells
-
total number of cells in the table.
- ngrp
-
the total number of patterns in the data.
- npatt
-
the number of groups of missing data.
- grouped
-
T if a frequency variable was input to the constructor. Oterwise, F.
- sj
-
the number of the last missingness patterns for which the jth variable
needs to be imputed to complete the monotone pattern.
- slevs
-
a list containing the names associated with each of the levels in each
column of
x
.
SEE ALSO:
,
,
,
,
.