Rev Language Reference
srGelmanRubin - Gelman–Rubin (PSRF) stopping rule
Allow an MCMC run to terminate once the specified criterion has been met.
The Gelman–Rubin rule compares the variance between runs with the variance within runs; its value tends to unity (1) as runs converge. It is widely referred to as the "potential scale reduction factor" (PSRF).
Usage
Arguments
| R : | RealPos (pass by value) |
| The maximum allowed potential scale reduction factor. | |
| filename : | String (pass by value) |
| The name of the file containing the samples. | |
| frequency : | Natural (pass by value) |
| The frequency how often to check for convergence. | |
| Default : 10000 | |
| burninMethod : | String (pass by value) |
| Which type of burnin method to use. | |
| Default : ESS | |
| Options : ESS|SEM |
Details
Because the statistic is defined by comparing the variation between different runs to the variance within each run, it can only be calculated when multiple independent runs are performed, by setting the `nruns` argument to `mcmc` or `mcmcmc` to a value greater than one.
Example
Binomial example: estimate success probability given 7 successes out of 20 trials
r ~ dnExp(10) p := Probability(ifelse(r < 1, r, 1)) n <- 20 k ~ dnBinomial(n, p) k.clamp(7) mymodel = model(k)
moves = VectorMoves() moves.append( mvSlide(r, delta=0.1, weight=1) )
paramFile = “parameters.log”
monitors = VectorMonitors() monitors.append( mnModel(filename=paramFile, printgen=100, p) )
Stop when the potential scale reduction factor falls below 1.01
stopping_rules[1] = srGelmanRubin(1.01, file = paramFile, freq = 1000)
Create the MCMC object
mymcmc = mcmc(mymodel, monitors, moves, nruns = 2)
Begin the MCMC run
mymcmc.run(rules = stopping_rules)
Edit this page