Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Arima in Stata vs. SAS


From   kristian@hefting.dk
To   statalist@hsphsun2.harvard.edu
Subject   st: Arima in Stata vs. SAS
Date   Mon, 6 Sep 2010 12:29:58 +0200

Hi

I have estimated the same sarimax-model in SAS and Stata.

I have used the estimates from SAS to forecast in Excel. These forecasts
are the same as the forecasts I obtain by using the forcast-statement in
proc arima.

Now I have estimated the same model using Stata. The parameters are more
or less the same (some small differences which I guess is due to SAS and
Stata using different maximization algorithms when maximizing the
log-likelihood function). But now I have entered the parameters from this
estimation into the Excel spreadsheet replacing the estimates from SAS to
obtain the forecasts. These forecasts are not the same as the ones I
obtain using predict in arima postestimation.

So I have the following two questions:
1. Shouldnt these statements (SAS and Stata) produce the same estimates?:

SAS
proc arima data=data;
identify var=SBP(1) crosscorr=(dw_INDO d_INDO);
estimate p=(2,5)(48) q=(1,2,6)(48)  input=(dw_INDO d_INDO) noconstant;
run;
quit;

Stata
arima D.SBP d_INDO dw_INDO, ar(2 5) ma(1 2 6) sarima(1,0,1,48) noconstant

2. If not is there any difference in how the two programs report the
parameters? (I know that the MA-part is reported as negative in Stata.)

Regards
Kristian Hefting


*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index