Fixed effects forecast
by Ricardo Gonçalves Silva
Hi,
Can Gretl forecast 1 to 3 periods ahead an estimated panel data(fixed-effects with no iterations) model?
HTH
RIck
13 years, 4 months
Paper advocating open source software and gretl
by Talha Yalta
Dear gretl users:
You might be interested to hear about my new paper entitled "Should
Economists Use Open Source Software for Doing Research?" published
this month in Computational Economics. The paper investigates
econometric software reliability and advocates the use of open source
software by taking gretl as a case study and showing how responsive
and transparent its development process is. I think many people here
might find it an interesting read.
More information and the download link is available here:
http://ideas.repec.org/a/kap/compec/v35y2010i4p371-394.html
I can send a working paper version if you do not have access to the above.
Cheers
A. Talha Yalta
--
“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)
--
13 years, 9 months
Re: [Gretl-users] introspection in gretl?
by Allin Cottrell
On Mon, 29 Nov 2010, Skipper Seabold wrote:
> I haven't been about to find much about this. Is there any
> (programmatic) way to quickly tell what is availble using the $
> operator after estimating a model? E.g., in R you can do
>
> mod <- lm(...)
> names(mod)
We now have something like this (in CVS and snapshots). I have
pressed the "varlist" command into service: if you give the
new --accessors option you get a list of the currently available
internal variables. For example
open data4-10
ols 1 0 2
varlist --accessors # or varlist -a
Allin Cottrell
14 years
Re: [Gretl-users] Using system, restrict, estimate within a function
by Allin Cottrell
On Mon, 29 Nov 2010, Allin Cottrell wrote:
> I'll aim to make the following variant on Ofer's function work:
>
> function ...
> system method=fiml
> # system spec
> end system
>
> restrict
> # restriction spec
> end restrict
>
> estimate method=fiml
> end function
It turns out this was simple to do. It's now in CVS and snapshots
(and will be in gretl 1.9.3, which should appear shortly).
Allin
14 years
Re: [Gretl-users] introspection in gretl?
by Allin Cottrell
On Mon, 29 Nov 2010, Skipper Seabold wrote:
> I haven't been about to find much about this. Is there any
> (programmatic) way to quickly tell what is availble using the $
> operator after estimating a model? E.g., in R you can do
>
> mod <- lm(...)
> names(mod)
No, at present there isn't a way to do this. But I can see it
might be quite helpful, and if it were programmed properly would
compensate for any out-of-dateness in the documentation for
the $ accessors.
Allin Cottrell
14 years
introspection in gretl?
by Skipper Seabold
I haven't been about to find much about this. Is there any
(programmatic) way to quickly tell what is availble using the $
operator after estimating a model? E.g., in R you can do
mod <- lm(...)
names(mod)
Thanks,
Skipper
14 years
Re: [Gretl-users] Using system, restrict, estimate within a function
by Allin Cottrell
On Mon, 29 Nov 2010, Sven Schreiber wrote:
> IIRC adding models to the model table (what "xxx" <- system would do) is
> not allowed inside a function, because functions are not supposed to
> affect/alter the outer scope. There was a discussion on this before I
> think (may have been on the devel list instead of the users list).
Yes, that's right: you can't save named models inside functions.
However, I see that this can be a problem.
We should be able to fix this by allowing "restrict" and
"estimate" to operate on the (anonymous) "last model" in the case
where the last model is a system. That is, I'll aim to make the
following variant on Ofer's function work:
function ...
system method=fiml
# system spec
end system
restrict
# restriction spec
end restrict
estimate method=fiml
end function
Allin Cottrell
14 years
Using system, restrict, estimate within a function
by Ofer Cornfeld
Hi,
I have tried to use system, restrict and estimate within a function in order
to perform restricted VAR, passing the series I want to estimate as
parameters as follows:
function varappx(series p_d,series d_d, series r, scalar rho)
"xxx" <- system
equation p_d 0 p_d(-1) d_d(-1) r(-1)
equation d_d 0 p_d(-1) d_d(-1) r(-1)
equation r 0 p_d(-1) d_d(-1) r(-1)
end system
restrict "xxx"
1 - rho*b[1,1] - b[2,1] + b[3,1] = 0
-b[2,2] + b[3,2] - rho*b[1,2] = 0
-b[2,3] + b[3,3] - rho*b[1,3] = 0
end restrict
estimate "xxx" method=fiml
end function
Gretl fails on the system command with "Syntax error in command line".
The same syntax outside the function works just fine.
What am I doing wrong?
With kind regards,
Ofer Cornfeld
TAU
14 years
accessing elements of series
by Bob McCall
I have a time series dataset and want to access a variable as a subscripted
array. Then I could do something like;
loop for i=1 .. 5
print v1[i]
endloop
but that doesn't work. I tried:
genr matrix X = v1
which gives me a Tx1 matrix equal to v1. I then tried:
print X[1,1]
but X is not the name of a variable or something. It didn't work. Can a
series be treated like a subscripted array?
Thanks,
bob
14 years, 1 month
Re: [Gretl-users] Gretl crashes when running a series of TSLS regressions via script
by Allin Cottrell
On Thu, 25 Nov 2010, Yangbo Du wrote:
> Whenever I run TSLS using the attached script file for each country's
> pure rate of time preference across three regressors (each of the three
> UN HDI components) and three instruments, gretl crashes after only a few
> regression runs...
Ah, the crash is coming from the lapack library, and the reason is
that we're asking it to do a QR decomposition on a matrix with
more columns than rows. The pattern of missing values in your data
is such that in your seventh model you have more instruments than
observations. This is fixed in CVS: instead of a crash you should
get an error message.
Allin Cottrell
14 years, 1 month
