Several minutes ago I run on
Ubuntu 15.04 the shell script below
./configure --enable-openmp --enable-build-doc --with-odbc
make -j 2
sudo make install
I get SVAR.pdf datet September 3, 2015
whereas in XP snapshot I have a newer
version of the same file, dated October 6, 2015
It seems I have missed something with ./configure.
Em 13 de outubro de 2015, Oleg escreveu:
What is TSS = RSS+ESS ?
> 1) residual sum of squares
> 2) regression sum of suares
> 1) explained sum of squares
> 2) error sum of squares
> Some people, e.g. R team, proposed
> to solve this enigma using MSS (model sum of squares)
> Does all this really important?
(this message was sent to the following thread:
Dear Allin and Jack,
The "ESS" abbreviation stands for "error sum of squares"?
1) I am afraid, I have missed the instruction.
What should I do with older packages?
Insert tags and re-upload them as soon as possible,
or I can wait till I'll upload upgraded versions?
2) What is supposed to be uploaded under
C88: Other Computer Software tag?
it wasn't easy hunting this one down in the sense of provoking the error
in a small (probably not minimal) example:
function scalar heyhey (list in)
var 20 in # 20 is too much -> error
if 5 < 10
list lulu = LRM LRY IBO IDE
catch scalar what = heyhey(lulu)
When I run this under Sep 20th's snapshot I get an error telling me I
have an unmatched "endif" ('ungepaartes "endif"' in German), which I
think is quite obviously not true.
I also would be grateful for any hints about workarounds, because I want
to use the real-world version of this type of code on a remote machine
where I cannot update the gretl version myself.
I have a very small complaint to make about the GUI label "Model" in the
main gretl window. I think Model is misnamed since it doesn't contain any
models. The drop down list consists of estimators, right? I make a big
deal about the differences between models and estimators in class, there it
is in gretl contradicting me.
Professor of Economics and Department Head
Economics and Legal Studies in Business
Oklahoma State University
I'm referring to tests after a 'var' command, perhaps (probably?)
similar problems affect other estimators:
1) modtest --normality --quiet: "--quiet" has no effect (everything is
2) modtest --normality --silent: "--silent" has no effect (everything is
3) modtest --autocorr --quiet: ok (same as silent, nothing)
4) modtest --autocorr --silent: ok (nothing)
(And BTW, there are only separate tests for each equation, don't we have
any Portmanteau type of test?)
5) modtest --arch --quiet: ok
6) modtest --arch --silent: some message stubs are still printed
This is still with 1.10.2.
As it is clear from [Patterson, K. D. Unit root tests in time series, volume 1, 13.7]
in the absence of seasonal unit roots, seasonality is just a particular case
of autocorrelation, which is routinely dealt with including sufficient lags
in ADF-type tests or sufficiently large window in PP-type tests.
So, with the first generations tests we have two possibilities:
include seasonal dummies, or set minimum lag length to ($pd-1)
With tests having no include seasonal dummies option the second
one is the only valid way to proceed.
So, my proposition is to include option:
set minimum lag length as ($pd-1) for zero frequency (i.e. usual,
non-seasonal) unit root tests