bdp_dw_mcmc: MH sampler for BDP-DW method

In beyondWhittle: Bayesian Spectral Inference for Time Series

MH sampler for BDP-DW method

Description

Obtain samples of the posterior of the Whittle likelihood in conjunction with a Bernstein-Dirichlet prior on the spectral density.

Usage

bdp_dw_mcmc(
  data,
  m,
  likelihood_thinning = 1,
  mcmc_params,
  prior_params,
  monitor = FALSE,
  print_interval = 100
)

Arguments

data	numeric vector.
m	window size needed to calculate moving periodogram.
likelihood_thinning	the thinning factor of the dynamic Whittle likelihood.
mcmc_params	a list generated by bdp_dw_mcmc_params_gen.
prior_params	a list generated by bdp_dw_prior_params_gen.
monitor	a Boolean value (default FALSE) indicating whether to display the real-time status
print_interval	If monitor = TRUE, then this value indicates the number of iterations after which a status is printed to console; If monitor = FALSE, it does not have any effect

Details

Further detail can be found in the simulation study section in the references papers.

Value

list containing the following fields:

k1, k2, tau, V, W1, W2	posterior traces of PSD parameters
tim	total run time
prior_params	the specifications of the prior

References

Y. Tang et al. (2023) Bayesian nonparametric spectral analysis of locally stationary processes ArXiv preprint <arXiv:2303.11561>

beyondWhittle documentation built on June 22, 2024, 11:35 a.m.