Metropolis-Hastings

From Christoph's Personal Wiki
Revision as of 00:11, 30 December 2005 by Christoph (Talk | contribs) (Started article)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Metropolis-Hastings (also known as Metropolis-Hastings-Green) is the main form of Markov chain Monte Carlo. The algorithm provides a probabilistic rejection rule to reject some proposed moves of an arbitrary irreducible Markov chain so that the resultant Markov chain has the desired stationary distribution.

Properties of the chain

  • If we find a tree with a better P(D|T), we are guaranteed to add that T to the chain.
  • If the new tree has a worse P(D|T), we may stay with the better tree, or switch. The chance of switching depends on the relative difference.
  • Less likely to include trees with low P(D|T) in chain
  • More likely to stay with a tree with high P(D|T) for many steps.
  • "Climbing mountains" but allowing some chance of moving downhill as well.

See also