Page 1 of 1
Seasonal Adjustment
Posted: Wed Oct 15, 2008 2:10 am
by MartinF
Hello,
I want to create an pseudo out of sample forecast program, but I only have non seasonal adj. data.
Well, I have monthly data of inflation and I want to have a new calibration of the seasonal component. So I just want to have monthly dummies like in these equation.
y=c(1)+c(2)*m2+c(3)*m3+c(4)*m4+c(5)*m5+c(6)*m6+c(7)*m7+c(8)*m8+c(9)*m9+c(10)*m10+c(11)*m11+c(12)*m12
with m2=1 if Februar and m2=0 if not Februar
and so on...
I don't know how I can create the seasonal dummies m2,m3...m12???
Can somebody help me?
Thanks
MartinF
Re: Seasonal Adjustment
Posted: Wed Oct 15, 2008 3:37 am
by startz
Hello,
I want to create an pseudo out of sample forecast program, but I only have non seasonal adj. data.
Well, I have monthly data of inflation and I want to have a new calibration of the seasonal component. So I just want to have monthly dummies like in these equation.
y=c(1)+c(2)*m2+c(3)*m3+c(4)*m4+c(5)*m5+c(6)*m6+c(7)*m7+c(8)*m8+c(9)*m9+c(10)*m10+c(11)*m11+c(12)*m12
with m2=1 if Februar and m2=0 if not Februar
and so on...
I don't know how I can create the seasonal dummies m2,m3...m12???
Can somebody help me?
Thanks
MartinF
Re: Seasonal Adjustment
Posted: Wed Oct 15, 2008 5:19 am
by MartinF
Thank you for your comment, but I have an additional question.
Is it possible to integrate the Census X12 application in a routine?
First I have a sample of 200 observations, then go back to 100.
Then I make an seasonal adjustment of the series by Census X12.
After that I estimate a 1 step forecast.
Well, now extend my range to 101.
Then SA by Census X12
1 step forecast.
range extention to 102
...
up to range 200
after that I want to compare different models by the RootMSFE.
Is it possible to integrate Census X12 by an command?
Best regards
MartinF
Re: Seasonal Adjustment
Posted: Wed Oct 15, 2008 5:30 am
by startz
Thank you for your comment, but I have an additional question.
Is it possible to integrate the Census X12 application in a routine?
First I have a sample of 200 observations, then go back to 100.
Then I make an seasonal adjustment of the series by Census X12.
After that I estimate a 1 step forecast.
Well, now extend my range to 101.
Then SA by Census X12
1 step forecast.
range extention to 102
...
up to range 200
after that I want to compare different models by the RootMSFE.
Is it possible to integrate Census X12 by an command?
Best regards
MartinF
but see the Command Reference for a long list of options