And be careful, in general, with missing data imputation. It should be used
sparingly. Parameter estimates for substituted data are generally unbiased,
but variance estimates are underestimated - it looks like you have more
data that your really do.
--
MORPHMET may be accessed via its
Andrea is correct. I am currently immersed in a complete graphics rewrite
facilitated by an enormous amount of initial work by Detelina Stoyanova.
She developed a generic data-side API to the nuts-and-bolts OpenGL
underpinning. I am currently integrating that into Morpheus, then must move
all
