My apologies if I have missed something, but I would like to ask
whether it is possible to extend genr in some way so as to provide
conditional definitions of variables. To give an illustration, in
Stata one can do the following:
gen ln_x = -10000 [where ln_x is a series]
replace ln_x = ln(x) if x > 0
This is very useful in a whole range of circumstances where one wants
to avoid the generation of missing values by handling the cases that
might generate them explicitly, but there are many other
circumstances where the syntax may be useful.
As far as I can work out, there are two ways of doing this in gretl
at present, both of them clumsy. One is by an explicit loop over
each of the elements of x with an if .. else .. endif. The other,
slightly more baroque method and not always possible, is to use
matrix functions combined with dummy variables.
The context in which I want to use something of this kind is in
writing a function to calculate residuals derived from a stochastic
frontier. In this case the calculation may fail because the divisor
is zero, but I know that the correct residual is 0 in such cases so I
would like to return zero rather than a missing value.
I use the syntax above very frequently when handling data in
Stata. As far as I understand from the manual Boolean constructs in
genr are restricted to the creation of dummy variables, so they do
not serve the purpose I have in mind.
Hi, are you sure that strings such as
should be marked for i18n?
It's the first time that this kind of strings appear in the .po file.
GPG/PGP Key-Id 0x943A5F0E - http://www.linux.it/~cri/cri.asc
for datasets with a lot of variables I think it would be useful if
'open' accepted a --quiet option, which would mean that the variable
names are not listed/printed.
what do you think?
cottrell(a)wfu.edu @ INTERNET skrev 2008-06-13 17:51:24 :
> On Fri, 13 Jun 2008, andreas.rosenblad(a)ltv.se wrote:
> > cottrell(a)wfu.edu @ INTERNET skrev 2008-06-11 15:55:14 :
> > > > Unless anyone has any show-stopping problems to report, I plan to
> > > release gretl 1.7.5 tomorrow, after doing an OS X build tonight.
> > >
> > > Allin.
> > > I think it looks great. Some comments:
> > > 1. The normtest command should be accessible also from the GUI.
> > E.g. Variable > Normality test
> Yes, that is planned, but I didn't want to introduce yet another new
> string for translation at this point.
> > 2. The quantile regression function: When the confidence >
> intervals cannot be computed (I get a "Not a number in >
> calculation" warning...
> Could you send an example where you get the "Not a number" message?
> It may be there's nothing we can do about that, but it might be
> worth investigating.
Open the Ramanathan 4-1 data set. Set price as dependent and sgft, bedrms,
baths as independent. Set desired quantiles to 0.25 0.50 0.75. Choose
"compute confidence intervals". Press OK: "Not a Number in calculation"
> > and I instead let gretl compute the standard errors for a tau >
> sequence, the graph with confidence intervals for the tau > sequence
> is still available, giving weird CIs.
> It's intended that the tau graph is still available -- the idea is
> that it will be based on +/- so many standard errors. I've
> experimented with this and got OK results. Could you post an
> example where the CIs are "weird"?
Take the same example as above but choose "compute standard errors"
instead. The graph of tau sequence for e.g. baths looks weird (very short
CI for 0.25 compared to 0.50 and 0.75). But maybe it isn't wrong, since I
see that the standard errors are much smaller for tau=0.25 than for 0.50 or
> > Further, for the quantile regression function the Gretl User?s >
> Guide says that for large data sets, "If you want confidence >
> intervals in such cases, you are advised not to use the > --
> intervals option, but to compute them using the method of > ?plus or
> minus so many standard errors?." It would be great if > there was an
> option to automatically compute and print the CI > calculated with
> this option. E.g. a new radio button in the GUI: > "Compute
> confidence intervals using normality approximation".
> OK, that's a possibility for the future. If you select just one tau
> value you can get such intervals by the usual means (Model window,
> Analysis, Confidence intervals), but I guess you're talking about
> the multiple-tau case.
I just thought that if the computations fail (got "Not a number" message)
or it takes too long time to compute the CIs, one could choose the
asymptotic method for the multiple-tau case.
Apologies for the repetition, but for anyone who's not on the
gretl-announce list: gretl 1.7.5 is now officially released.
Thanks to Sven, Andreas and Marcin for prompting some last-minute
Department of Economics
Wake Forest University, NC
cottrell(a)wfu.edu @ INTERNET skrev 2008-06-11 15:55:14 :
> Unless anyone has any show-stopping problems to report, I plan to
> release gretl 1.7.5 tomorrow, after doing an OS X build tonight.
I think it looks great. Some comments:
1. The normtest command should be accessible also from the GUI. E.g.
Variable > Normality test
2. The quantile regression function: When the confidence intervals cannot
be computed (I get a "Not a number in calculation" warning, and I instead
let gretl compute the standard errors for a tau sequence, the graph with
confidence intervals for the tau sequence is still available, giving weird
Further, for the quantile regression function the Gretl User’s Guide says
that for large data sets, "If you want confidence intervals in such cases,
you are advised not to use the --intervals option, but to compute them
using the method of “plus or minus so many standard errors”." It would be
great if there was an option to automatically compute and print the CI
calculated with this option. E.g. a new radio button in the GUI: "Compute
confidence intervals using normality approximation".
Best regards and thanks for a GREAT software,
I wrote, in response to Sven on apparently interminable printing:
> I just created an artificial dataset with 250 observations on
> 550 variables, opened it, and issued the "print" command in the
> console. Yes, it took about 40 seconds of 100% CPU, but then I
> got the printout OK.
> This is mostly a GUI performance issue. The same thing in
> gretlcli executes in just a second or two.
Aha, I've found the performance bottleneck on this point.
I had thought it was GTK, but in fact it was a gross inefficiency
in gretl's handling of printing to an in-memory buffer. Printing
of large chunks of data in the GUI should now be very much faster.
(It's more or less instantaneous for the 250 x 550 dataset.)
I have a couple of questions that came up when trying to get the
(German) translation ready for 1.7.5:
1) %g\\%% confidence interval -- is the double backslash here really ok?
2) output stream -- what's the context / what does it mean?
3) scatterplot with control -- is "control" short for "control
variable"? (I think there is also another case of this usage somewhere else)
4) /Graphs/tau sequence -- what does it mean?
5) Creation and I/O -- sounds almost religious, again what does it mean?
6) Stickiness... -- this refers to script output windows right? But I
can't think of anything good, any German suggestions out there?