On Mon, 9 May 2011, Bob McCall wrote:
> When calculating information criteria for arima models, Gretl
> seems to use k + 1 parameters instead of k parameters.
Gretl calculates the information criteria for ARIMA models in the
same way as X-12-ARIMA. The "extra" parameter is the variance of
On Tue, 17 May 2011, Daniela Tatiane dos Santos wrote:
> I estimated a model with quadratic and linear adjustment, using
> consumption as a function of income per capita. However, I would
> like the adjustment was cubic.
Then just add a cubic term?
x_cubed = x^3
We can't have a menu item for each and every mathematical
transformation: use /Add/Define new variable, or use the console,
or a script.
My name is Daniela. I wonder how can I create a panel model using the cubic
form. The gretl has the ability to create a quadratic model, but I do not know
where is the icon for the cubic model.
If I could create a command, using the model in panel (stand and random effects)
what would the command to the cubic model?
Daniela Tatiane dos Santos
I am pleased to announce that volume 41 of Journal of Statistical Software
(out today: http://www.jstatsoft.org/v41) contains an article written by
yours truly on state-space models in gretl. It should be viewed as
complementary to the manual chapter on Kalman filtering, and contains two
Before you say "Hey, since when it's ok to spam the list with your own
papers?", let me just say that I honestly think it can be useful to the
medium-to-advanced gretl user, I really do. Of course, comments are
Riccardo (Jack) Lucchetti
Dipartimento di Economia
Università Politecnica delle Marche
On Tue, 10 May 2011, cociuba wrote:
> I was thinking if we will see Gretl on Ipad/Android platform in the near
I have no plans of that sort, but if anyone wants to try building
gretl for those platforms I'll be willing to help with any changes
to the build files that might be required.
On Tue, 10 May 2011, Henrique Andrade wrote:
> Suppose we have two series named "teste_1" and "teste_2".
> If someone try to make X-12 adjustments through the GUI
> ("Variable -> X-12-ARIMA analysis"), Gretl overwrites the
> seasonally adjusted series due to the fact that it uses the
> same file name for the adjusted series "test_d11".
In CVS I've now adjusted things so that up to 8 characters of the
original series name are retained in the name of the saved series
from x12a. This is enough to fix the particular problem noted. A
more rigorous solution would be to get user input on the naming of
the output series but that will require some more work.
sorry for posting such a probably off-topic question to this list but I
don't know of any other place I could ask.
I am fitting an arima model with exogenous variables and was trying to
identify if there exist differences between the relationship depending on
the season of the year.
Therefore I added a combined Variable which combines a binary dummy for the
season (Summer, Fall, Winter) with the exogenous variable the logarithmic
average temperature on a particular day. So I have three Dummies (LogAvgTemp
x Summer, ..., LogAvgTemp x Winter).
Now I don't know if it's sufficient if I add those three combined dummies or
if I should also add the LogAvgTemp as an additional exogenous variable.
Perhaps someone could point me to a text book or something similar which
might address such a problem.
Thanks in advance and once again sorry for being slightly off-topic,
Dear Gretl Developers,
Today a colleague of mine noticed that Gretl's X-12-Arima
analysis has an important problem.
Suppose we have two series named "teste_1" and "teste_2".
If someone try to make X-12 adjustments through the GUI
("Variable -> X-12-ARIMA analysis"), Gretl overwrites the
seasonally adjusted series due to the fact that it uses the
same file name for the adjusted series "test_d11".
Moreover, Gretl maintains the same file description:
"seasonally adjusted teste_1 (X-12-ARIMA)"
Henrique C. de Andrade
Doutorando em Economia Aplicada
Universidade Federal do Rio Grande do Sul