“What methods can be used to speed up the processing time when using rstanarm for joint modeling to make population-level (i.e., marginal) predictions with approximately 200,000 longitudinal data points and over 10,000 survival data points?”
Hi, @qqq and welcome to the Stan forums.
I’m afraid there’s not much user flexibility in how to speed up rstanarm processing time. It’s possible that the model you’re trying to fit could be coded directly in Stan for more efficiency, but the rstanarm models tend to be pretty optimal.
If the model’s relatively easy to fit, but just has a lot of data, you can consider reducing the number of warmup and sampling iterations. I’m not sure how much rstanarm allows you to control initialization, but some kind of reasonable starting point can help.