Markov Chain Monte Carlo: Sampling from your posterior distribution

The idea behind MCMC spans decades of research from some of the smartest people. If you’re delving into this literature, its sometimes good to visualize what this method is meant for. Here is a visualization of the delayed rejection adaptive metropolis MCMC method sampling from a “banana distribution”. The pdf is shown in gray, the accepted points have the red x and the accepted ones are retained on the figure.