CARR model
by Burak Korkusuz
Hi,
When I run the code below for the CARR model, gives the error saying smpl bob ; and does not generate forecasts.I could not understand why that s happening. Though, when I change fixed window with increasing window; i.e. the line "smpl 1+i 1000+i" with "smpl 1 1000+i" it does work.Does anyone have know why that s happening?thanks,
include carr.gfnopen http://ricardo.ecn.wfu.edu/pub/gretldata/2005-CARR.gdt
# --- estimate and forecast ECARR(1,1)set verbose offseries pointforecast = NAloop i=1..100 -q smpl 1+i 1000+icarr(R,1,1,null,1)fcast 1001+i 1001+ipointforecast[1001+i] = $fcast endloop
error in new starting obs*** error in function carr, line 187> smpl bob ;>> carr(R,1,1,null,1)
2 years, 6 months
Prediction intervals vs confidence intervals
by Brian Revell
Can one assume that the interval in the forecast GUI Analysis option of an
estimated model is a true prediction interval?. If indeed it is, it would
also be useful to be able to graph the prediction interval surrounding the
fitted function values as well as the forecast ones.
Brian J Revell
Professor Emeritus
Harper Adams University , Shropshire
Current Chair of Defra Economic Advisory Panel
Tel 01952 815237
Tel: +44 1952 728153
Mbl +44 7976 538712
University: +44 1952 815235
alt: email: bjrevell(a)harper-adams.ac.uk
2 years, 6 months
Loops and user-defined functions in mle command
by Alecos Papadopoulos
Good morning.
I have a likelihood computed by applying quadrature twice, in a
double-sum way. I can compute the inner sum and then use an index loop
to compute the outer sum and the whole likelihood.
My questions are
a) Can I include a loop inside an mle command? My impression is, "no".
b) If I write the computation of the likelihood in a user-defined
function, can I include this function in an mle command?
If the answer is again "no", than I guess I will have to, reversely,
include the mle command in a subsequent user-defined function, as shown
in the user's guide, right?
--
Alecos Papadopoulos PhD
Athens University of Economics and Business
web: alecospapadopoulos.wordpress.com/
scholar:https://g.co/kgs/BqH2YU
2 years, 6 months
gretl 2021d released
by Allin Cottrell
As usual, see http://gretl.sourceforge.net/
This is mostly a bug-fix release, but there are two new features to
report:
* An enhancement of the "biprobit" command, whereby rho, the
correlation between the errors in the two equations, is available as
part of the parameter vector, $coeff, and the covariance matrix,
$vcv.
* Support for linear programming, courtesy of the lpsolve library.
On Linux you need to install the lpsolve package to activate this
feature (but that's easy via your distro's package manager); in our
Windows and macOS packages we supply a suitable build of the
library. PDF documentation can be found under the gretl Help menu
("Linear programming").
Here's the change log:
2021-09-30 version 2021d
- "biprobit" command: include rho in $coeff, $stderr and
$vcv
- New function lpsolve(): supports linear programming via
the lpsolve library
- Prevent opening of general and console preferences
dialogs simultaneously
- Fix partial breakage of the panel-time plotting facility
- Fix potential failure mode in echoing script input lines
in the GUI program
- Fix a small and somewhat unlikely memory leak
- Fix a GUI build error when using GTK 2
- Fix Windows bug: greek-letter identifiers not working
in some locales
- Fix Mac bug: binary data read/write not accessible
- Fix Linux bug: possible crash when running MPI from the
GUI program
Allin
2 years, 6 months
