Generate Matrix Time Series following MAR(p)

This function generates MAR(p) time series array.

Usage

sim_mar(
  num_sim,
  num_burn = floor(num_sim/2),
  p = 1,
  row_coef,
  col_coef,
  row_sig = diag(ncol(row_coef)),
  col_sig = diag(ncol(col_coef)),
  init = array(0L, dim = c(ncol(row_coef), ncol(col_coef), p))
)

Arguments

num_sim

Number to generated process

num_burn

Number of burn-in

p

MAR lag

row_coef

Row MAR coefficient (np x n).

col_coef

Column MAR coefficient (kp x k).

row_sig

Row covariance of innovation matrix. By default, identity matrix.

col_sig

Column covariance of innovation matrix. By default, identity matrix.

init

Initial \(Y_1, \ldots, Y_p\) array or list to simulate MAR model. By default, zero.

Value

n x k x (num_sim - num_burn) array.