Simulate distribution from fit

Modeling
1

I have fit a model with my in-sample data. I am trying to simulate a distribution of predictions for each row in my hold out data set.

So far I found the posterior_predict function, but this returns simulations of the in sample data updated from the new data. This is not what I am looking for.

I am looking for something analogous to applying the simulate() function with a fit object from glm()

How do I simulate a full distribution from each row of new data?

2

Hi. Perhaps I am misunderstanding your question, but have you used the newdata = and summary = FALSE arguments in predict? That should get you what you’re asking for (sample draws from the posterior predictive distribution for your holdout dataset).