Subscribe to DSC Newsletter

I am using SAS/ETS for my forecasting task in my current job but I find it not impressive.

Some people told me that there are some other packages that are better suited for time series analysis and forecasting such as RATS, Eviews, SCA, etc.
Can anyone make suggestion on their choice of software for forecasting? Thanks.

Views: 95

Replies to This Discussion

STATA has a good set of time series tools, I use it for and studies of clinical trials and economic data. They have a great set of documentation too. I bet it would be useful for forecasting.
Thanks Elaine! Have you used SAS/ETS for your studies? If you have, could you tell me the advantages of STATA over SAS/ETS based on your experience? Thanks a lot.
I think SAS does have enough in it for the Time Series. You do have a range of options from ARIMA to UCM to STATESPACE models available. I think if you can really explore your data more, this should come in handy. Besides, whatever software you may use, the statistics is going to be the same. So I believe you should be able to work your way around it!
I guess the answer would depend on why you find SAS to not be impressive.

SAS has a fairly good time series package that I used in time series classes at two schools… I now use Stata, and strongly prefer it as my general stats software. I’d also say it has a slight edge over SAS as far as time series are concerned, though SAS has improved dramatically in recent years in the area.

Other packages… RATS is widely used in academic macroeconomist circles, but I think SAS has surpassed it (or at least caught up to it) in recent years as far as featureset is concerned. I’d stay far far away from Eviews. Never used SCA, never known anyone to use SCA… I used TSP as a master’s student and can’t say anything exciting about it. My first time series professor swore up and down that BMDP has the most stable convergence algorithms. I played around with JMulTi a year or two ago. It’s free and has a lot of features for economic data, but I wouldn’t call it a full-featured time series package.

If I was a SAS user, I doubt I’d be terribly interested in learning another package to do time series…


On Data Science Central

© 2020 is a subsidiary and dedicated channel of Data Science Central LLC   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service