Hi all, I am tyring to do some simulations with Metropolis Hastings. The proposal distribution is student t distribution and the stationary distribution is inverse gamma. However, there is a logical error here that IG distribution requires positive values whereas numbers are generated from stationary distribution which gives both positive and negative values? This is all explained in the article called "Classification of RNA-Seq Data via Gaussian Copulas"

I will appreciate if anyone can help me to sort this problem out? Thanks in advance.

Necla

Similar questions and discussions