MCMC (12): A real example in Bayesian model selection

In the previous example, we don’t really need a MH method, since we know it is normal distribution. MH is really useful when we don’t know the close form of our target distribution. Let’s look at a real example in Bayesian model selection.