#147 StrucTiSM package: Add exogenous variables into univariate time series basic structural model forecast

[Brian J Revell]

The STS BSM can be estimated incorporating exogenous variables, but they cannot be incorporated into the forecast function option (evidently it would only be relevant if projected values of these were to be user provided). This seems to be a significant limitation of the current BSM version.

[Brian J Revell]

The present BSM is limited in usefulness for projection purpises.

[Sven Schreiber]

Thanks for the reminder. Yes it would be useful. (Internal note: Perhaps it would make sense to leverage the ksimul apparatus for that.)

[Sven Schreiber]

That feature is implemented in a not-yet-released version of the package. We are still trying to verify results with some known-good cross-check from some other software. Hopefully, that shouldn't take too long.

[Brian J Revell]

Hi Sven
Much appreciated
Brian

On Tue, 9 Nov 2021, 07:10 Sven S., svetosch@users.sourceforge.net wrote:

• status: open --> closed
• Comment:

Version 0.7 of the package has been released with this feature, so I'm
closing this.

---

• [feature-requests:#147]
  https://sourceforge.net/p/gretl/feature-requests/147/ StrucTiSM package:
  Add exogenous variables into univariate time series basic structural model
  forecast*

Status: closed
Group:
Created: Sat Mar 27, 2021 11:47 AM UTC by Brian J Revell
Last Updated: Sat Oct 09, 2021 07:04 PM UTC
Owner: nobody

The STS BSM can be estimated incorporating exogenous variables, but they
cannot be incorporated into the forecast function option (evidently it
would only be relevant if projected values of these were to be user
provided). This seems to be a significant limitation of the current BSM
version.

---

Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/gretl/feature-requests/147/

To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/

[Sven Schreiber]

Version 0.7 of the package has been released with this feature, so I'm closing this.

[Brian J Revell]

Is there a simple way to deselect specific Panel units in the Panel
Regression Sample menu?
For example I have a Panel of 12 units. Whilst I can delete say units 1-3
by resetting the sample to 4-12, or say units 9-12 by setting the sample to
units 1-8 I cannot for example remove units 6 and 8, as the sample
selction menu does no lend itself to that option.
It would be helpful to be able to have an option to itemise the specific
panel units in the sample menuu,either to include eg 1,2,3,4,5,7,9,10,
11,12, or to exclude 6 and 8..
Suggestions welcomed

Brian J Revell
Harper Adams University UK

[Sven Schreiber]

Brian, you're replying to a ticket that has been closed for over two years, and the topic is unrelated. If the problem persists after my reply, please re-post your question (or follow-ups) on the mailing list.
Having said that, you would have to choose Sample/Restrict by condition, and then enter an expression like "unit != 6 && unit != 8" (without the quotes). If the series "unit" doesn't exist yet, create it with "Add/Panel group index" first (or similar wording).
Whether that qualifies as a "simple way", I don't know.

[Brian J Revell]

Is there any reason why recursive forecasts cannot be included in the ARIMA
option for univariate modelling together with confidence intervals when no
exogenous varaibles are included in the specification Clearly post sample
data the MA terms would drop out.

Brian

Brian J Revell
Professor Emeritus (Agricultural Economics)
Harper Adams University UK
Current Chair of Defra Economic Advisory Panel
Former President Agricultural Economics Society

Tel: home 01952 728153 Mobile 07976 538712
Address: Orchard Croft, Vineyard Rd, Homer, Much Wenlock TF!3 6NG
Alt. Email bjrevell@harper-adams.ac.uk

[Sven Schreiber]

Hi, again I don't think that you're actually talking about the topic in this (closed) ticket, which was about StrucTiSM. Please re-post your question on the gretl mailing list. (like you did with your other message today)

[Brian J Revell]

O k
Though the issue was one of a feature request. I will selete the feature
request address from my contacts.
B

On Sat, 23 Nov 2024, 16:26 Sven S., svetosch@users.sourceforge.net wrote:

Hi, again I don't think that you're actually talking about the topic in
this (closed) ticket, which was about StrucTiSM. Please re-post your
question on the gretl mailing list. (like you did with your other message
today)

---

[feature-requests:#147]
https://sourceforge.net/p/gretl/feature-requests/147/ StrucTiSM package:
Add exogenous variables into univariate time series basic structural model
forecast

Status: closed
Group:
Created: Sat Mar 27, 2021 11:47 AM UTC by Brian J Revell
Last Updated: Tue Feb 06, 2024 03:45 PM UTC
Owner: nobody

The STS BSM can be estimated incorporating exogenous variables, but they
cannot be incorporated into the forecast function option (evidently it
would only be relevant if projected values of these were to be user
provided). This seems to be a significant limitation of the current BSM
version.

---

Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/gretl/feature-requests/147/

To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/