Hi all!
I was wondering if it is possible to create an empty
series (like a column array), estimate a model, then run a fcast on the sample and… make the
empty array including the forecasted values. Something like this:
series Forecast = NA
smpl
1 1000
logit Y 0 X –-p-values
fcast 1 1000
Forecast = fcast ?
How can I create the forecast array in order to
generate a new time series under Y and X which contains the forecasted values?
I have another question directly linked to the first
one: is it possible to use fcast
in a rolling sample? Something like this:
smpl
1 200
loop for i = 1 1000
logit Y 0 X –-p-values
fcast 1+i 200+i ?
smpl +1 +1
endloop
Is there anyone who can help me? Reading the manual I
discovered that the –-rolling option
for fcast
is only available with OLS estimate, but with a lot of model (Logit for
example, or correlated uhat models) we usually use MLE.
Thank you so much!
Oscar