create_rand_coef_mat function

Creates a random coefficient matrix

create_rand_coef_mat(
  k,
  p,
  max_abs_eigval = 0.8,
  sparsity_pattern = c("none", "lasso", "hvar"),
  sparsity_options = NULL,
  decay = 0.5,
  ...
)

Arguments

• k: Number of time series

• p: Number of lags

• max_abs_eigval: if < 1, then the VAR will be stable

• sparsity_pattern: The sparsity pattern that should be simulated. Options are: "none" for a dense VAR, "lasso" for a VAR with random zeroes, and "hvar" for an elementwise hierarchical sparsity pattern

• sparsity_options: Named list of additional options for when sparsity pattern is lasso or hvar. For lasso the option num_zero

  determines the number of zeros. For hvar, the options zero_min (zero_max) give the minimum (maximum) of zeroes for each variable in each equation, and the option zeroes_in_self (boolean) determines if any of the coefficients of a variable on itself should be zero.

• decay: How fast should coefficients shrink when the lag increases.

• ...: Not currently used

Returns

Returns a coefficient matrix in companion form of dimension kpxkp.