HOW TO GET H-STEP AHEAD FORECASTS

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

sofie hunt
Posts: 3
Joined: Wed Jun 25, 2014 2:49 pm

HOW TO GET H-STEP AHEAD FORECASTS

Postby sofie hunt » Wed Jun 25, 2014 2:59 pm

Hi,

Please I am very new to E-views but luckily I have learnt a lot. I would like to know how to generate h-step ahead forecasts for a recursive window. i.e 1980q1 to 2010q4 to estimate 2011q1 and then 1980q1 to 2011q1 to estimate 2011q2. Mainly to get the Mean squared error
I wouldn't mind having the commands, I am using both eviews 7 and 8 but mostly 7.

Also how to do the same using a VAR model along with the Mean squared error also.

And lastly how to do the Diebold and Mariano test on eviews, as I do not seem to find it.

Thank you.

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13604
Joined: Tue Sep 16, 2008 5:38 pm

Re: HOW TO GET H-STEP AHEAD FORECASTS

Postby EViews Gareth » Wed Jun 25, 2014 3:13 pm

None of the things you describe are built in to EViews, you'll have to program them yourself.

http://forums.eviews.com/viewtopic.php? ... 848#p14848

sofie hunt
Posts: 3
Joined: Wed Jun 25, 2014 2:49 pm

Re: HOW TO GET H-STEP AHEAD FORECASTS

Postby sofie hunt » Thu Jul 10, 2014 8:50 pm

thank you so much for your response.

Please how do I use principal components to generate pseudo out of sample forecasts...is it possible to use principal components to generate forecasts on eviews. If yes please how can I do it.

EViews Glenn
EViews Developer
Posts: 2682
Joined: Wed Oct 15, 2008 9:17 am

Re: HOW TO GET H-STEP AHEAD FORECASTS

Postby EViews Glenn » Fri Jul 11, 2014 10:00 am

Do you mean principal components scores for the observed or forecasts?

sofie hunt
Posts: 3
Joined: Wed Jun 25, 2014 2:49 pm

Re: HOW TO GET H-STEP AHEAD FORECASTS

Postby sofie hunt » Sat Jul 12, 2014 2:26 am

I guess so because I am supposed to apply a forecast combination technique using principal components......

what I have done so far is to check the significant Eigen value and determine how many principal components to use (in which case was mostly one even though I had 8 variables) then I run a regression on the dependent variable with the 1 PC to get my forecast series and error

However I am not sure of this procedure


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 2 guests