EViews.com

Not that I can think of. To be honest though, you might have been better off not using the add-in at all, and just writing your own routine to do the ARMA selection, using the code here as a starting point. Then you have much more control over what is reported, kept, etc.... Yeah, you're probably r...

EViews Gareth wrote:I see nothing particularly wrong with what you're doing.

Yeah, it works for 19 series so it can't be wrong. But I'm wondering if there is something in the ARIMAsel code that could be preventing it from creating strings after a while

Thanks for your help, Gareth. I've got it working now but something weird is happening. I'm running a look for 54 variables but it is stopping at the 19th one. The reason is no string seems to be getting created after the 19th series. I'm not sure why this is the case. My code follows. My 54 series ...

But I'm using data that has been transformed the way I want and then setting diff=0 so that isn't an issue!

Sorry, I meant suppose my series is called GDP, how can I have it be called GDP in the string as well instead of being assigned the temp name. This may sound trivial but I need to run a loop that estimates 54 equations based on the criteria chosen by ARIMAsel for each of 54 variables.

Thanks. It is now doing what I want it to but the string being saved looks like this:

_TEMPSER01 C MA(1)

Obviously there is no series called _TEMPSER01. Is there something I can alter in the program code for _TEMPSER01 to be the actual series name?

Gareth, I am running ARIMAsel through 54 series. I would like to set up a program that runs ARIMAsel on all 54 series and then stores equation objects for each. Then I can use these for forecasting etc. I'm not sure how to go from having a string with the best specification to having an equation be ...

Gareth, I would like the ARIMAsel function to store the selected equation when it estimates it. I added the line of code you suggested to another poster but it doesn't see to do the trick. string chosenstr = {%eqname}.@spec Any idea what the issue could be? Or is that not quite what the above is sup...

Thanks for your response, Glen. With a bit of work I can see how I could perhaps construct it on a case-specific basis. Was wondering if there was a more general way already incorporated in EViews that I was not aware of. At this point for my purposes it seems I may be able to get by without program...

anyone?

EViews Gareth wrote:I don't follow what you're asking.

I think he wants to do pseudo out of sample forecasting but the method he proposes to do so seems strange to me. Rather than messing around with the database you can just set up your equations to forecast recursively.

A rolling forecast is similar to a recursive forecast but different in that the recursive forecast uses an expanding window instead of a rolling window. However, if someone has written a program that does rolling forecasts it should be very easy to change it a bit to do recursive forecasts instead....

Any VAR(p) process can be written in its companion form as a VAR(1).

Is there a way of obtaining the companion coefficient matrix?

I would appreciate a response, as I require this for thesis work.

A rolling forecast is similar to a recursive forecast but different in that the recursive forecast uses an expanding window instead of a rolling window. However, if someone has written a program that does rolling forecasts it should be very easy to change it a bit to do recursive forecasts instead.

I increased the number of observations and it seems to get closer, so that must be it. Thanks for this program. It is extremely useful.