this is weird
by Summers, Peter
Folks,
I'm estimating a series of regressions with varying break dates. I set up a script with the break dates in a vector, then generate a post-break dummy variable inside a loop, like so:
<script>
matrix breaks = breaks = {1985; 1985; 1986; 1986; 1982; ...}
loop ii=4..29
post = (obs>=obsnum(breaks[$ii-3]))
(run regression, do other stuff)
end loop
</script>
This worked in an earlier version of gretl (I wrote it a year or so ago), but now the "post" series returns all NA's. I've tried forcing breaks[$ii-3] to be an integer, but no change. I've played around a bit in the console window, and things work fine if I use "post = (obs>=obsnum(1985))," but not otherwise. For example, trying
tmp = breaks[$ii-3]
post = (obs>=obsnum(tmp))
doesn't work either. Any thoughts?
TIA,
PS
===============================
Dr. Peter Summers
Assistant Professor
Department of Economics
Texas Tech University
===============================
16 years, 2 months
Forecast ARIMA
by Александр Рогальский
Good day,
During the simulation the following problem occured.
This problem concerns the construction of confidence intervals.
Having developed the ARIMA model (Time Series->ARIMA, using the Exact LikelyHood method) I made a forecast of the process in а five-year term.
The predicted values were calculated correctly, but all confidence intervals had the same size (but the must expand with
increasing period of pre-emption).
What is the reason of it?
Alexander
16 years, 2 months
non-interactive R mode does not work but interactive mode does
by Ricardo Mayer
Hi everyone,
I have the following problem.
If I create a New R Script in Gretl, I'm able to run it using the
Interactive Mode, this is, R GUI opens successfully and does the
intended computation. But when try to run the same script using the
non-interactive mode, it pops an error message saying: "the system
can't find the specified path".
Considerations:
1) I'm running Gretl 1.9.1 and R 2.11.1 under Windows XP (but I had
the same problem on Windows Vista at home). This is the first time
I've tried Gretl, thus never had case where it worked, to compare.
2) This is the script I tried:
a <- 4
a
3) I checked the Tools -> Preferences -> General -> Programs and the
locations of Rgui.exe (which I guess matters for the interactive mode)
and R.dll (ditto for non-interactive mode, right??) both locations
are correctly specified. Just in case I clicked on Browse, find each
file and double clicked on them and pressed Apply.
This may be a well-known problem, but my search on gretl-users list
didn´t turned any hits for this issue.
thanks in advance,
Ricardo
16 years, 2 months
Saving Graphs
by Henrique Andrade
Dear Gretl Community,
I'm trying to save a correlogram graph with the Arial font (editing the line
# set term pngcairo font "sans,8"). Gretl shows me the correlogram with the
Arial font with no problems, but when I try to export to a PNG file the
Arial font is substituted by Sans. What can I do to solve this problem?
This same behavior is observed in both Windows XP and Mac OS/X.
Best,
--
Henrique C. de Andrade
Doutorando em Economia Aplicada
Universidade Federal do Rio Grande do Sul
www.ufrgs.br/ppge
16 years, 2 months
Re: [Gretl-users] LW estimator and the GPH test
by Javier García
Hi, everybody;
LW and GPH estimators are consistent for d=<1 and asymptotically normal for d<0.75 (see Velasco, 1999a, 1999b, Phillips and Shimotsu, 2004 and Phillips, 2007), but LW is more efficient. Besides, the asymptotic variance depends on the bandwidth (and, normally, the bigger the sample size the bigger the bandwidth). If you want to test the hypothesis of unit root you should apply the test on the first differences of the series and test if the memory parameter is equal zero. If you don't reject that hypothesis, the series seems to have a unit root. To avoid differenciation, there are other posibilities: tapering (Velasco, 1999a, 1999b), ELW of Shimotsu and Phillips (2005), non-local estimators (Abadir et Al., 2007), etc.
Cheers
Javi
References:
Velasco, C., 1999a. Gaussian semiparametric estimation of non-stationary time series. J. Time Ser. Anal. 20, 87-127.
Velasco, C., 1999b. ‘Non-Stationary Log-Periodogram Regression’, Journal of Econometrics, 91, 325-371
Phillips, P.C.B., Shimotsu, K. (2004). Local Whittle estimation in nonstationary and unit root cases. Ann. Stat. 32, 656-692.
PHILLIPS, P.C.B. (2007): Unit root log periodogram regression. Journal of Econometrics 138(1), 104-124.
ABADIR, K.M., DISTASO, W. and GIRAITIS, L. (2007): Nonstationary-extended local Whittle estimation. Journal of Econometrics 141, 1353-1384.
SHIMOTSU, K. and PHILLIPS, P.C.B (2005): Exact local Whittle estimation of fractional integration. The Annals of Statistics 33(4), 1890-1933.
16 years, 2 months
ARMAX problem
by Александр Рогальский
Good day,
I am using GRETL 1.8.7 for modeling dynamic econometric models. And
there is one problem that i can't make with it.
I don't know precisely how a model ARMAX is computed.
I use Model->Time Series->ARIMA for this set of data
Dependant variable: i_osn_sred
Independent variable: time
AR Order: 1
Difference: 0
MA order: 0
No seasonal data
Also i include a constant and check the box "Parameter covariance matrix
via Hessian".
The method is "Exact Maximum Likelyhood".
After the operations been completed I have a set of coefficients:
const 332395
phi_1 0.760190
time -7065.29
But i don't know precisely how a final equation looks like? Is this
right or not: y(t) = const + 0.76*y(t-1)-7064.29*t?
According to that equation there is a difference between values y(t)
and Analysis->Display actual, fitted, residuals.
How does this method work (ARMAX))? And Can I make myself a forecast
using this equation?
Thanks in advance.
Alexander
16 years, 2 months
sub-sampling problem
by artur bala
Hi allin,
I was running the following script and gretl reports an error I don't actually understand.
gretl version 1.9.1cvs
Current session: 2010-08-10 15:52
? open "C:\Program Files\gretl\data\misc\hamilton.gdt"
Read datafile C:\Program Files\gretl\data\misc\hamilton.gdt
periodicity: 12, maxobs: 202
observations range: 1973:01-1989:10
Listing 4 variables:
0) const 1) PC6IT 2) PZUNEW 3) EXRITL
? smpl ; 1973:12 --restrict
Full data range: 1973:01 - 1989:10 (n = 202)
Current sample: 1973:01 - 1973:12 (n = 12)
? series d1 = 1
Generated series d1 (ID 4)
? smpl 1974:01 1974:12 --restrict --replace
Variable number 13 is out of bounds
Error executing script: halting
I tried to reset the sample (with the --full flag) before restricting but obtained the same error.
cheers,
artur
Une messagerie gratuite, garantie à vie et des services en plus, ça vous tente ?
Je crée ma boîte mail www.laposte.net
16 years, 2 months
compilation error on linux
by Artur T.
Hi,
I just tried to compile gretl using the cvs source on my Ubuntu 10.04
system. Usually it works fine but today I got the following error after
the "make" command:
------------
............... fertig.
rm -f es.gmo && /usr/bin/msgfmt -c --statistics -o es.gmo es.po
es.po:1253:2: syntax error
es.po:1253: Schlüsselwort »es« ist unbekannt
/usr/bin/msgfmt: es sind 2 fatale Fehler aufgetreten
make[1]: *** [es.gmo] Fehler 1
make[1]: Verlasse Verzeichnis '/home/artur/cvs/gretl/po'
make: *** [po] Fehler 2
-------------
I already tried "make clean" but it did not solve the problem. I don't
know whether it is a bug or I did a mistake at some point.
Best,
Artur
16 years, 2 months
LW estimator and the GPH test
by Talha Yalta
Hi all,
Running the fractional integration tests on my data gives estimated
degree of integration = 0.92 and 0.94 for LW and GPH tests
respectively. I was wondering if these tests are suitable for near
unit root testing and whether I can interpret these results as
evidence of a near unit root? I couldn't find in the user's guide a
reference for these tests.
Cheers
Talha
--
“Remember not only to say the right thing in the right place, but far
more difficult still, to leave unsaid the wrong thing at the tempting
moment.” - Benjamin Franklin (1706-1790)
--
16 years, 2 months
Deleting variables
by Henrique Andrade
Dear Gretl Community,
I have 100 series inside my series (a .gdt file) and I would like to delete
all the series that I don't need. I know how to do this using the GUI or
with the command "delete", but I would like to make it in a more automatic
way.
For example, I would like to use the series 5, 8, 37 and, 98 and delete
anyone else. Please take a look in this pseudo-code:
list Y = 5 8 37 98
if the variable is on the list
make nothing
else
delete variable
endif
Best regards,
--
Henrique C. de Andrade
Doutorando em Economia Aplicada
Universidade Federal do Rio Grande do Sul
www.ufrgs.br/ppge
16 years, 2 months
