The help files do not make this clear, but predict and Mean in rms do not
support computation of predicted means for the continuation ratio model.
The help file for cr.setup shows you how to compute the unconditional
probabilities you would need for doing that.
Frank
apeer wrote:
>
> Dear list,
Dear list,
I'm currently trying to use the rms package to get predicted ordinal
responses from a conditional ratio model. As you will see below, my
model seems to fit well to the data, however, I'm having trouble
getting predicted mean (or fitted) ordinal response values using the
predict functio
2 matches
Mail list logo