I am doing some stochastic simulations with eviews7, with the attached code.
The code generate 20 new pages in which are stored all the repetitions for the 3 equations in the system over a forecasting sample of 60 periods.
I have several doubts:
1) the series generate are panel data and are structured as a column with the forecasting sample repeated 20000 times, can i reshape it in order to have series that look like matrix in which in each column i have a repetition?
2) Over this big series i need to work only with the forecast at 1step, 12steps and 60steps, can i deleate all the not needed solutions?
I wrote a code that made it, with matrix operator, but it does not work well because it generate to much vector, matrix and object and after a bit it go out of memory.
Thanks
Sfarz
Code: Select all
'Program estimate a VAR(1) with Full Infomation Maximum Likelihood and generate 20000 stochastic solutions for 60steps with bootstrap(same sample of estimation)
for !i=1 to 20
smpl 1972M01 1994M11+!i
sysvar1.fiml
smpl 1994M12+!i 1999M11+!i
modelvar1.update
modelvar1.stochastic(i=b,s=1972M01 1994M11+!i,r=20000,p=solutionvar1_!i)
modelvar1.solve(s=m)
next
