Help using Base SAS procedures

Proc Autoreg to model and then to forecast

Reply
N/A
Posts: 0

Proc Autoreg to model and then to forecast

Hi mates,

I am using the procedure autoreg to model a garch model and then use that model to do 1-step ahead forecast.
The codes look like:

proc autoreg data=dataset;
where date<='25dec2007'd;
model return / method=ml garch (p=1,q=1);
output out=outset predicted=p cev=v;

The autoreg procedures help in forming the fitted model, and give predicted value aside the all observations before 25dec2007. However, the dataset does not include any predicted value after that day.

In stead of 3,4,5-step forecasting, I would like to do 1-step forecast everytime when new data is observed beyond 25dec2007. The predicted model is the one found under autoreg procedures.

I actually could write the codes, according to the model's parameters, for generating the predicted value, but I wanna know if there is faster way to do the same in SAS. Is there any procedure useful for this modelling?

Thanks.

~ Patrick
N/A
Posts: 0

Re: Proc Autoreg to model and then to forecast

Posted in reply to deleted_user
Hi mates,

I am using the procedure autoreg to model a garch model and then use that model to do 1-step ahead forecast.
The codes look like:

proc autoreg data=dataset;
where date > .......;
model return / method=ml garch (p=1,q=1);
output out=outset predicted=p cev=v;

The autoreg procedures help in forming the fitted model, and give predicted value aside the all observations before 25dec2007. However, the dataset does not include any predicted value after that day.

In stead of 3,4,5-step forecasting, I would like to do 1-step forecast everytime when new data is observed beyond 25dec2007. The predicted model is the one found under autoreg procedures.

I actually could write the codes, according to the model's parameters, for generating the predicted value, but I wanna know if there is faster way to do the same in SAS. Is there any procedure useful for this modelling?

Thanks.

~ Patrick
Ask a Question
Discussion stats
  • 1 reply
  • 416 views
  • 0 likes
  • 1 in conversation