Fitting Bayesian MAR — mar

This function fits Bayesian Matrix Autoregressive (BMAR) model with various priors.

Usage

mar_bayes(
  y,
  p = 1,
  exogen = NULL,
  s = 0,
  factor_spec = set_matfactor(),
  num_chains = 1,
  num_iter = 1000,
  num_burn = floor(num_iter/2),
  thinning = 1,
  row_spec = set_mar_minnesota(),
  col_spec = row_spec,
  exogen_row_spec = row_spec,
  exogen_col_spec = row_spec,
  factor_row_spec = row_spec,
  factor_col_spec = col_spec,
  verbose = FALSE,
  num_thread = 1
)

Arguments

y: Matrix-valued time series data
p: VAR lag (Default: 1)
exogen: Unmodeled matrices
s: Lag of exogeneous matrices in MARX(p, s). By default, s = 0.
factor_spec: Augmented factor matrix specification.
num_chains: Number of MCMC chains
num_iter: MCMC iteration number
num_burn: Number of burn-in (warm-up). Half of the iteration is the default choice.
thinning: Thinning every thinning-th iteration
row_spec: Row coefficient specification
col_spec: Column coefficient specification
exogen_row_spec: Exogenous row coefficient prior specification.
exogen_col_spec: Exogenous column coefficient prior specification.
factor_row_spec: Factor row coefficient prior specification.
factor_col_spec: Factor column coefficient prior specification.
verbose: Progress log
num_thread: Number of threads

References

Chan, J. C. C. & Qi, Y. (2025). Large Bayesian matrix autoregressions. Journal of Econometrics, 105955.

Zhang, W. (2025). Bayesian Dynamic Factor Models for High-Dimensional Matrix-Valued Time Series. SSRN Electronic Journal.