Right now we use bounded Cauchy for both MAP and NUTS: <code class="

related warnings from stan <div class="snippet-clipboard-content notranslate posit

Reparameterization of sigma about orbit HOT 2 CLOSED

uber commented on July 27, 2024

Reparameterization of sigma

from orbit.

Comments (2)

edwinnglabs commented on July 27, 2024

related warnings from stan

Informational Message: The current Metropolis proposal is about to be rejected because of the following issue:
Exception: student_t_lpdf: Scale parameter is 0, but must be > 0!  (in 'unknown file name' at line 195)

If this warning occurs sporadically, such as for highly constrained variable types like covariance matrices, then the sampler is fine,
but if this warning occurs often then your model may be either severely ill-conditioned or misspecified.

Informational Message: The current Metropolis proposal is about to be rejected because of the following issue:
Exception: student_t_lpdf: Scale parameter is 0, but must be > 0!  (in 'unknown file name' at line 195)

If this warning occurs sporadically, such as for highly constrained variable types like covariance matrices, then the sampler is fine,
but if this warning occurs often then your model may be either severely ill-conditioned or misspecified.

from orbit.

edwinnglabs commented on July 27, 2024

I did some finding that the exact re-parametrization runs faster in mcmc and should be working in theory for mcmc. However, based on past experience, MAP may not be able to adapt that. In other words, if we want to implement this, we need to split the case of MAP and MCMC

from orbit.

Recommend Projects