Modelling a multivariate skewed t-distribution

Hey - I need to model / draw a vector of innovations from a multivariate skewed t-distribution proposed by Sahu et al (2003) (http://onlinelibrary.wiley.com/doi/10.2307/3316064/abstract) - the expression at equation 11. Any suggestions how to do this in Stan?

Best Regards
Rasmus

I cannot answer your question but someone posted a very similar question a few days ago:

1 Like

Maybe with the algebra solver in Stan. Here’s something that might help outside Stan.

2 Likes

That link is paywalled.

If you can define the log density, then you should be able to use it as a prior or likelihood in Stan. Stan lets you define new densities. You have to be careful with linear algebra stablity (e.g., use A / B rather than A * inv(B), for example) and try not to repeat calculations that can be reused.