Reversible-jump Markov chain Monte Carlo
(Redirected from Reversible Jump)
In computational statistics, reversible-jump Markov chain Monte Carlo is an extension to standard Markov chain Monte Carlo (MCMC) methodology that allows simulation of the posterior distribution on spaces of varying dimensions. Thus, the simulation is possible even if the number of parameters in the model is not known.
be a model indicator and
. The model indication need not be finite. The stationary distribution is the joint posterior distribution of
The proposal
, where
. The move to state
The function
must be one to one and differentiable, and have a non-zero support:
so that there exists an inverse function
that is differentiable. Therefore, the
is met where
. This is known as dimension matching.
The acceptance probability will be given by