Convergence problems with poor chain mixing pPCA model

Thank you for your quick reply and insights!

@ricardoV94
Yes, Z variables unfortunately go hand in hand with the observations because we are meant to have one per unit.
We are meant to truly know near to nothing about Z, so I resisted lowering its sigma, I might have to look into it.

I got right on to re-parametrizing W and it does seem to help a bit with the chain’s profile. Thank you!

@ckrapu and @jessegrabowski thank you! i’ll adapt some of those suggestions to Aesara.