acf and pacf output
by Allin Cottrell
I recently got an email from Alvaro Novo, who pointed out that the
ACF and PACF figures produced by gretl's "corrgm" command were not in
agreement with other packages.
I have now tested this myself (against gnu R and Eviews), and Alvaro
is right. I believe the gretl figures are asymptotically equivalent
to those given by the other programs, but since the gretl figures are
"non-standard" I have fixed this for the next release: in gretl 1.2.5
the ACF and PACF output will be in agreement with R.
In case anyone is interested in the details:
1. Gretl's figure for the k-order autocorrelation coefficient was
calculated as the correlation between y_t and y_{t-k}, which is OK,
except that the formula used treated y_t and y_{t-k} as if they were
distinct variables with (possibly) distinct means. The correct
formula uses a common value for the mean of y. In this case the old
gretl result would be asymptotically correct for a series that is
stationary in mean.
2. Gretl's PACF numbers used the estimator that is set out in several
standard sources (Greene, Hamilton, Davidson and MacKinnon): namely,
the coefficient for lag k is the last coefficient in an OLS regression
of y_t on [1, y_{t-1}, y_{t-2}, ..., y_{t-k}]. However, it appears
that the procedure used by R and Eviews is the Durbin-Levinson
algorithm, which computes the sample PACF recursively from the sample
ACF values. I have now switched to Durbin-Levinson. (From my reading
on this topic, I'm not quite clear on why the results differ: any
insights into that?)
--
Allin Cottrell
Department of Economics
Wake Forest University, NC
20 years, 8 months
- 3
- 3
details on gretl 1.2.4
by Ignacio Diaz-Emparanza
Allin,
First of all, I would like to transmit my congratulations for the new features
of gretl, I think they are new steps in the way to obtain a very complete
GNU econometric program.
I would also like to report some small problems with this new version:
(This time I am running gretl-1.2.4 in Spanish with Linux Mandrake 10.0)
- I had libpng.so.3 installed but the installation of gretl had a dependency
problem with libpng.so.2, so finally I had to install a package from another
distribution (different from Mandrake) having exactly the library
libpng.so.2.
- The accents appear correctly in the graphs (using gnuplot installed trough
your RPM file), and saving the graph in a file as .EPS they are shown
correctly, but if I save two or more graphs in the "session-graphic page",
the accents are not displayed.
- In generating this "session-graphic page" with several plots, I think dvips
should be invoqued with the option "-o '%S.ps'" , because I have just
installed Mandrake 10.0 and the default option for dvips was sending the file
to the default printer. I had to change the config.ps file in my Tetex
distribution to do this "graphic-page" work correctly.
-Untranslated string: "convergence achieved after 10 iterations" appears in
ARMAX and GARCH.
--
Ignacio Díaz-Emparanza
Departamento de Econometría y Estadística
Universidad del País Vasco/EHU.
http://www.bl.ehu.es/~etpdihei/
20 years, 9 months
- 2
- 1
- ← Newer
- 1
- Older →