sim.VAR.novol: Simulate data from Gaussian BVAR model without SV

In hoanguc3m/fatBVARS: Bayesian VAR with Stochastic volatility and fat tails

Simulate data from Gaussian BVAR model without SV

Description

This function simulate data from a BVAR model without SV.

y_t = B x_t + SQRT(w_t) A^(-1) Sigma eps_t

Usage

sim.VAR.novol(
  dist,
  K = 5,
  p = 2,
  t_max = 1000,
  b0 = 0.5,
  a0 = 0.5,
  h = 0,
  nu = 6,
  gamma = 0.5,
  y0 = matrix(0, ncol = K, nrow = p),
  sigma_G = NULL,
  seednum = 0,
  burn_in = 0
)

Arguments

dist	The variable specifies the BVAR error distribution. It should be one of c("Gaussian","Student","Skew.Student","Skew.Student", "MT","Skew.MT","MST").
K	The number of variables in BVAR model.
p	The number of lags in BVAR model.
t_max	The number of observations
b0	The coefficients of B matrix.
a0	The coefficients of A matrix.
h	The log diag(Sigma) matrix.
nu	The degree of freedom.
gamma	The skewness from [-gamma, gamma].
seednum	The default seed.
burn_in	The discarded observations.

Value

A list of simulated data with its specification.

Examples

## Not run: 
datagen <- sim.VAR.novol(dist="Gaussian")
y <- datagen$y

## End(Not run)

hoanguc3m/fatBVARS documentation built on Jan. 12, 2023, 4:42 p.m.