Multi-step ahead forecast

[Nicole]

Do I maybe have an error in the last line of my "period spesification?"

21-period-ahead forecast
%21pers=@otod(@dtoo(%start)+!i+!window-1)
%21pere=@otod(@dtoo(%start)+!i+!window+20)

or should I change how many steps the forecast should move at a time?

'move sample !step obs at a time
for !i=1 to !lenght-!window+1-!step step !step
!j=!j+1

[ying]

hello, Nicole.
I have exactly met the same problem. I have tried hard but can't figure out.
Would you please tell me how you solved it at last ?
Thank you very much.


ying

[walmir.silva]

Hi guys,

I would like to model a regime switching with a rolling window of 1000 days, my problem is how to store the regime probabilities for each one of the iterations in the loop. I know that with
eq01.makergmprobs(type=smooth) rprob1 rprob2
I can store but only for the last iteration.

I'm adapting a code posted in this forum Multi-step ahead forecast by EViews Gareth on Jan 29, 2014.

'rolling Markov Switching regression
group y variable

' set window size
!window = 1000

' declare equation for estimation
equation eq01

' get size of workfile
!length = @obsrange

'calculate number of rolls
!nrolls = @floor(!length-!window)

'matrix to store coefficient estimates
matrix(5,!nrolls) coefmat

'matrix to store regime probabilities
matrix(1000,!nrolls) regime_probabilities

' loop
for !i = 1 to !nrolls

' dynamic sample
smpl @first+!i @first+!i+!window

equation eq01.switchreg(type=markov) y c
eq01.rgmprobs(type=smooth) 1 2
' estimate equation -

colplace(coefmat,eq01.@coefs,!j) 'store coefficients

eq01.makergmprobs(type=smooth) rprob1 rprob2

colplace(regime_probabilities,eq01@..... ????? ‘ how to store the regime probabilities rprob1 e rprob2????
'in this point, I would like to store rprob1 e rprob2 in matrix regime_probabilities for each loop

next


Thanks for any help,

Walmir