I adapted the dd_hyperbolic function in hBayesDM to fit behavioural data in RStan 2.21.0.
There seems to be no problem with model specification as it runs normally, and there are no divergencies or any other convergence issues. However, when checking the summary of the Stanfit object, I seem to have NaNs on some Rhat values of the y pred, as well as on some SE values of the y pred as well. Please, see example of .csv attached.
example.csv (405 Bytes)
How can I interpret these NaNs values?
I’m new to Stan and I’ve fitted another model before but never encountered this issue.
Thank you for your help!