Enables sampling from arbitrary distributions if the log density is known up to a constant; a common situation in the context of Bayesian inference. The implemented sampling algorithm was proposed by Vihola (2012) and achieves often a high efficiency by tuning the proposal distributions to a user defined acceptance rate.
Andreas Scheidegger, andreas.scheidegger@eawag.ch or scheidegger.a@gmail.com
| Package: | adaptMCMC |
| Type: | Package |
| Version: | 1.4 |
| Date: | 2021-03-29 |
| License: | GPL (>= 2) |
| LazyLoad: | yes |
The workhorse function is MCMC. Chains can be updated with
MCMC.add.samples. MCMC.parallel is a
wrapper to generate independent chains on several CPU's in parallel
using parallel. coda-functions can be used after conversion
with convert.to.coda.
Vihola, M. (2012) Robust adaptive Metropolis algorithm with coerced acceptance rate. Statistics and Computing, 22(5), 997-1008. doi:10.1007/s11222-011-9269-5.
MCMC, MCMC.add.samples,
MCMC.parallel, convert.to.coda