On Fri, 14 Jun 2013, Gabriela Nodari wrote:
I would like to obtain simulated data by bootstrapping the residuals
of a
VAR model. I have gave a look to the gretl guide, and I found something
related to resampling and bootstrapping (section 7.4) [...]
Check out the varsimul() function. Here's a simple example of
use:
<hansl>
open data9-7
var 4 PRIME UNEMP --silent
matrix A = $compan[1:2,]
# residuals + const
matrix U = $uhat .+ $coeff[1,]
# initial values
matrix y0 = {PRIME, UNEMP}[1:4,]
# simulate the VAR
matrix X = varsimul(A, U, y0)
# check that we got back the original data
# (to within machine precision)
eval X - {PRIME, UNEMP}
# now resample U
matrix Us = resample(U)
matrix Xs = varsimul(A, Us, y0)
# look at the simulated data
print Xs
</hansl>
Use "set seed" if you want to be able to replicate the
resampling. Put the last few lines in a loop for multiple
simulations.
Allin Cottrell