Access latent space in MV families

Hi all,
I am developing a model like this.

\begin{aligned} Y &\sim N(\mu_Y, \sigma_Y)\\ Z &\sim N(\mu_Z, \sigma_Z)\\ \mu_Y &= X^T\alpha\\ \mu_Z &= \mu_Y \times \gamma + X^T\beta \end{aligned}

Is there a way to access the latent dparam of Y (\mu_Y) in current brms? Or any trick to do so?

If I understand correctly you’re trying to use \mu_Y across two outcomes’ formulas; that is not yet possible but is planned for brms v3.

Thanks. Good to hear that it is in the plan. Do you have any timeline for that?

If I had the skills to contribute to v3 trust me I would… This is up to the devs (so basically Prof. Bürkner).

1 Like