Suppose we decide to run an MCMC to sample from the posterior of a binomial proportion theta. They use the MCMCpack package and run the MCMCmetrop1R() function, with a burnin = 1000, starting the chain at 0.5.
What does the burnin = 1000 mean?

Question

Suppose we decide to run an MCMC to sample from the posterior of a binomial proportion theta. They use the MCMCpack package and run the MCMCmetrop1R() function, with a burnin = 1000, starting the chain at 0.5.

What does the burnin = 1000 mean?

Answer

Only every 1000th draw is kept from the chain, which removes the correlation between neighbouring values in the output.

Answer

The first 1000 iterations are used to tune the proposal, and the chain is then restarted from 0.5 for the 5000 draws.

Answer

The first 1000 iterations are run but not stored, so the starting value of 0.5 has less influence on the draws we keep.

Answer

An extra 1000 iterations are run after the 5000 draws, to check that the chain has settled on the target distribution.

Answer

The first 1000 iterations are used to adjust the prior, so that it agrees better with the 31 successes in 45 trials.

Crowdly

Suppose we decide to run an MCMC to sample from the posterior of a binomial prop...

Want instant access to all verified answers on learning.monash.edu?