state series of a kalman filter

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

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

state series of a kalman filter

Postby lisa » Thu Oct 27, 2011 2:21 pm

hello,

I have estimated a VAR model whith time varying parameters using the kalman filter with the eviews software. the model contains 4 variables, so the there are 16 states. I want to extract the state series. I wrote 5 series in the place reserved in the series_names and i can't add the athors (there is no place to add them) and when I tryed to obtain 4 or 3 states each time, eviews don't give me any result. please who knows how can I obtain the 16 state series.
Please I'm very pressed with time and i need a solution.

Thanks a lot.

EViews Glenn
EViews Developer
Posts: 2672
Joined: Wed Oct 15, 2008 9:17 am

Re: state series of a kalman filter

Postby EViews Glenn » Thu Oct 27, 2011 3:46 pm

Looks like the dialog isn't accepting data beyond a certain point. We'll get it fixed.

For now, you can either (1) use the wildcard get generate all 16 with the wildcard expression (which apparently everyone does), or (2) use the command line.

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: state series of a kalman filter

Postby lisa » Thu Oct 27, 2011 4:03 pm

Please how can I use the command line? I don't understand.

EViews Glenn
EViews Developer
Posts: 2672
Joined: Wed Oct 15, 2008 9:17 am

Re: state series of a kalman filter

Postby EViews Glenn » Thu Oct 27, 2011 4:11 pm

You can enter the command

ssname.makestates [list of names]

where ssname is the name of your state space object, in the white area at the top of your EViews screen. Note that if you want something other than the default one-step ahead predicted states, you'll have to use the t= option, as in

ssname.makestates(t=smooth) [list of names]

to get the smoothed states

If you aren't really comfortable with using commands, the easiest thing to do is to just leave the "*f" in the output edit field of the dialog (or change it to some other wildcard expression). EViews will make all of your states with the name of the state replacing the wildcard (in the default case with "*f" it will make state series with the names of the states followed by the letter "f"). You can rename the saved states after your are done...

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: state series of a kalman filter

Postby lisa » Thu Oct 27, 2011 4:40 pm

thank you

EViews Glenn
EViews Developer
Posts: 2672
Joined: Wed Oct 15, 2008 9:17 am

Re: state series of a kalman filter

Postby EViews Glenn » Wed Nov 02, 2011 8:26 am

A fix for the dialog issue has gone in the main code base and should appear in the next EViews patch.

lisa
Posts: 45
Joined: Thu Jul 28, 2011 3:27 pm

Re: state series of a kalman filter

Postby lisa » Sun Nov 27, 2011 12:45 pm

hello,
my problem is not solved.
when I wrote " ss01.makestates (t=filter) [sv11 sv12 sv13 sv14 sv21 sv22 sv23 sv24]" , eviews don't give me any result.
please help me.
thanks again.

startz
Non-normality and collinearity are NOT problems!
Posts: 3775
Joined: Wed Sep 17, 2008 2:25 pm

Re: state series of a kalman filter

Postby startz » Sun Nov 27, 2011 1:06 pm

lisa wrote:hello,
my problem is not solved.
when I wrote " ss01.makestates (t=filter) [sv11 sv12 sv13 sv14 sv21 sv22 sv23 sv24]" , eviews don't give me any result.
please help me.
thanks again.

If you have a space between makestates and the option that might be a problem.

EViews Glenn
EViews Developer
Posts: 2672
Joined: Wed Oct 15, 2008 9:17 am

Re: state series of a kalman filter

Postby EViews Glenn » Thu Dec 01, 2011 10:44 am

And there shouldn't be square brackets. Those were just to indicate syntax.


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 47 guests