Gretl 1.6.0 review
DownloadGretl is a cross-platform software package for econometric analysis, written in the C programming language
|
|
Gretl is a cross-platform software package for econometric analysis, written in the C programming language. It is is free, open-source software.
You may redistribute it and/or modify it under the terms of the GNU General Public License (GPL) as published by the Free Software Foundation.
Here are some key features of "Gretl":
Easy intuitive interface (now in French, Italian, Spanish, Polish and German as well as English)
A wide variety of least-squares based estimators, including two-stage least squares and nonlinear least squares
Single commands to launch things like augmented Dickey-Fuller test, Chow test for structural stability, Vector Autoregressions, ARMA estimation
Output models as LaTeX files, in tabular or equation format
Integrated scripting language: enter commands either via the gui or via script
Command loop structure for Monte Carlo simulations and iterative estimation procedures
GUI controller for fine-tuning Gnuplot graphs
Link to GNU R for further data analysis
Requirements:
gtk
gtkextra (optional)
libxml
lapack
gnuplot
gdk-pixbuf
readline (optional)
gnome (optional)
libole2 (optional)
GMP (optional)
What's New in This Release:
- Revamp mechanism for saving sessions in gui program (we
now save sessions as zipfiles containing XML representations
of any models that have been estimated)
- Add "native" exact ML estimation of ARMA models using the
Kalman filter
- Saving data in native gretl format: make gzip compression the
default
- Add some basic functionality for analysis of discrete variables.
New behavior of frequency distribution, plus new commands:
"discrete" (interpret variables as discrete), "dummify" (create a
set of dummies coding for the values of a discrete variable), and
"xtab" (cross-tabulation of two discrete variables)
- Don't add an extra newline at the end of plain text data files
(e.g. CSV)
- "restrict" command for single equations: add a printout of the
restricted coefficient estimates and standard errors
- Add mechanism for packaging user-defined functions so as to
make these available for other users; also add a mechanism
for downloading such packages from a server, and for uploading
new packages to the server
- Redefine the weight variable in gretl's "wls" command, for
better compatibility with other programs: the weight var is
now defined as the series of diagonal elements of the weighting
matrix, so existing gretl scripts that use wls should square
the old weight variable
- Don't print the "model has no constant" message, and do not
use special methods to calculate R^2 and F, if the set of
regressors spans the space of a constant
- Indexing of coefficients in "restrict" command: standardize on
using a 1-based index in all contexts
- Revamp panel data estimation: redefine the "panel" command for
fixed- or random-effects estimation, rename the old "panel"
command as "paneldat"; refine calculation of Hausman test statistic
- Lags in main window: make these visible under their parent
variable, via a "twisty" triangle
- Revamp multiple-precision models: add more statistics, and make
them available for capture via genr
- Revamp test for differences of two means in the GUI hypothesis
test calculator
- Imposing panel structure on a data set: add the option of
defining the structure by reference to two index variables, one
for the unit and one for the period -- this is implemented in
the "setobs" command and the GUI data structure wizard
- Add --robust option to nls command (nonlinear least squares)
- Add --hessian and --robust options to the mle command, which
respectively produce a covariance matrix based on the Hessian
and a QML sandwich estimator of the covariance matrix
- Add Quandt likelihood ratio (QLR) test for structural break
- ADF and KPSS tests: add option to difference the selected
variable before testing
- Delimiter for "CSV" files: add semicolon to the list of
options
- VECMs: make Johansen's alpha and beta retrievable as matrices
with identifiers $jalpha and $jbeta
- Several bug-fixes for the matrix manipulation facility, plus
a few new functions (meanr, meanc)
- GUI menus restructured and polished
- Add facility for multiple time-series graphs from the GUI
- Revamp the "--ten" option for printing variables to 10 digits:
replace with a configurable "--long" option
- Graph editing dialog: make the line width configurable, and
also make the preferred line width settable per variable in
the "Edit attributes" dialog
- Data import and export: add option for reading and saving data in
JMulTi format
- Databases: add capacity to read PcGive in7/bn7 databases
- Gretl command minibuffer: model commands now produce GUI
model windows (for single-equation methods)
- QR decomposition option: fix the F-statistic values produced
for autoregressive models
- pvalue() function: always return the right-hand value, but add
a complementary cdf() function that always gives the left-hand
value
- NIST linear regression test suite: re-express accuracy in terms
of log relative error, plus other enhancements
- Generating random variables: add option to generate t and
chi-square values in GUI; add option of setting the PRNG seed
to the GUI variable-generation dialogs
- Time-series filters: Add Hodrick-Prescott and Baxter-King filters
as a GUI menu item, along with simple moving average and
exponential moving average
- Engle-Granger cointegration test: improve printout, add option
to test down for the lag order, and fix an issue with the
selection of the MacKinnon p-value for this test
- Tobit: enable handling of missing observations for this estimator
Gretl 1.6.0 keywords