Martin, thanks.
However, I like to compare the results for different models.
I used the regression and the var(vector autoregression). I have in my
mind to use arima too.
My time variable takes values from 1 to 80.
As far as regression, in order to forecast the values I executed the
following commands:
matrix coeff=e(b)
matrix colnames coeff=L.bs L2.bs L3.bs L4.bs L5.bs L6.bs L7.bs L8.bs
L9.bs L10.bs L11.bs L12.bs L13.bs L14.bs L15.bs L16.bs L17.bs L18.bs
L19.bs L20.bs L21.bs L22.bs a asqr fb mr ar my jn jl ag sp ot nv dc L.cs
const
tsset t
tsappend, add(18)
tsset t
predict bs if t<=81
predict cs if t<=81
gen const = 1
matrix score bs=coeff if t>=81, replace
list t y bs if t>=81
I am getting an empty table!
I am not sure with the command "matrix colnames coeff" as far as the
monthly dummy variables and quadratic time trend
Also, with the "predict" command I am not sure if I have to predict the
monthly dummy vars and quadratic time trend since those are given.
Thank you
-----Original Message-----
From: Demetriou, Eleftherios
Sent: Friday, August 14, 2009 8:11 AM
To: '[email protected]'
Subject: Forecast values and get their interval extrapolation forecasts
Hi,
I regress y on L(1/22).y t tsquare feb-dec x_1
Those are time series in monthly.
I would like to obtain a forecast for future y up to 18 periods ahead
with their interval.
Any help with the commands I should use would be much appreciated.
Thank you
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
