Page 1 of 1

Simulating Random Walk / White Noise Processes

Posted: Sun Oct 26, 2008 12:26 pm
by nchesley
Can anyone share some insight into simulating a random walk process with distribution properties similar to the observed distribution of a time series variable?

Context: I have tested several ARMA(p,q) and GARCH specifications on a series of monthly price data and want to compare the output against a random walk process that has a behavior similar to the observed series.

Thanks,
NC
:eviews6:

Re: Simulating Random Walk / White Noise Processes

Posted: Sun Oct 26, 2008 2:13 pm
by startz
Can anyone share some insight into simulating a random walk process with distribution properties similar to the observed distribution of a time series variable?

Context: I have tested several ARMA(p,q) and GARCH specifications on a series of monthly price data and want to compare the output against a random walk process that has a behavior similar to the observed series.

Thanks,
NC
:eviews6:
if the observed series is X, use the commands

Code: Select all

smpl @all series simx = 0 smpl 2 @last simx = @mean(d(x),"@all") + simx(-1) + @stddev(d(x),"@all")*nrnd