Open
Description
Hey,
I can't figure out an easy way to use a Turing model produced using turing_model
to predict on some hold-out data. In Turing one usually passes missing
for the predicted values. However, this seems to not be possible in the turing_model
step. What workflow do you recommend? I would like to avoid building a second equivalent model which uses the sampled parameter values as this somehow defeats the purpose of this package.
Thanks in advance,
Heiner
Metadata
Metadata
Assignees
Labels
No labels