Syntax
where S is a belief store number and
The VARDATA: command is used to generate a variance-covariance
matrix in a belief store from the sample variances and covariances
calculated from data attached to the corresponding elements.
In the first form of the syntax, sample covariances are calculated, and
sample correlations are calculated for the second form of the syntax.
It is assumed that each of the elements comprising the bases B1 and B2
possesses data, and that there exists a subset of cases for which data
is available for all of these elements. This subset defines the number
of cases used to define the sample covariances (or correlations). Use of
the correlation option requires that all the respective variances be
positive. We use the multiplier
are the names
of bases.
rather than
when determining these summary statistics.
The beliefs generated are checked for symmetry (see §6.8.5).
,B 
) 