CODEWITHSUNDEEP
bayesianmcmcdirichlet
LifeisGood94
LifeisGood94

Reputation: 67

What does the Hastings Ratio look like for MCMC with a Dirichlet Process Prior?

So I've been having a tough time understanding this. This paper by Neal did the best to further my understanding of using a Dirichlet process prior, but I'm still not fully understanding.

Let's say my data is modeled via a Gaussian Mixture Model (GMM). If I proposed the components w with a beta(1,epsilon), and then I put a prior of gamma(1,1) on epsilon, what should the numerator of the Hastings ratio look like? The paper talked about using the conditional distribution of epsilon which is where I get stuck.

Thanks in advance.

Upvotes: 0

Views: 23

Answers (0)

Related Questions