I am trying to use the S-Plus library, MIX (by Schafer) to create
multiple data sets (the original data set has missing values). I have
a problem in understanding the use of the "margins" vector which has
to be submitted to the program when using the restricted general
location model.
I "seem" to understand its use when operating under logistic or
proportional odds type models (as they have a dependent variable).
Given that I supply the data matrix and and design matrix, the same
way I do in GLM and have no "margins" vector. Is there any way one can
explain the use of this particular vector approaching the problem as
if I have no missing values.
Thanks.
Vumani Dlamini
Swaziland
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at
http://jse.stat.ncsu.edu/
=================================================================
