It worked! Thank you very much, Max! I wasn’t aware of affine transformation before. One more question though, does the affine transformation differ from code the non-centered parameterization manually?
Edit: sorry for asking before research, I actually found this: Affine transformation in Stan
Follow-up: I tried both methods. Unfortunately, the affine transformation didn’t work as well as manually coded non-centered parameterization in terms of the number of effective samples. However, with the same seed, their estimations are very alike. Any explanation?
Affine transformation:
Manually coded NCP:
