Home  /  Stata News  /  Vol 35 No 1  /  Bayesian inference using multiple Markov chains
The Stata News

«Back to main page

In the spotlight: Bayesian inference using multiple Markov chains

Bayesian models have a lot to offer researchers, but the core computational tool of these models—Markov chain Monte Carlo—can be slow and is susceptible to nonconvergence. And when the Markov chain does not converge, you do not want to rely on the results for inference.

Want a formal diagnostic for Markov chain convergence?

Want to run your Bayesian analysis with multiple chains faster, perhaps orders of magnitude faster?

Interested in prosocial cooperative behavior of chimpanzees?

Read my colleague Nikolay's blog post.

— Vince Wiggins
Vice President and Head of Development