Undefined symbol get_gretl_rc_path
by Gordon Hughes
I have just installed Ubuntu 9.10 on a old computer. I downloaded
the current version of the gretl source and, after the usual process
of installing missing libraries, it compiled without any
problems. However, when I try to run gretl I get an error message
saying "symbol lookup error: /usr/bin/gretl_x11: undefined symbol:
get_gretl_rc_path.
My understanding is that gretl creates the file gretl2rc. when it
first runs. I have ensured that there is not an old version of this
file hanging around. What else do I need to do? Incidentally I have
checked that this does not happen when I install the earlier version
of gretl (1.8.1) from the Ubuntu repository, but I need to use the
current version to tidy up some of my functions before re-releasing them.
Gordon
14 years, 5 months
correlogram
by Javier García
Thanks, but I wanted to say is that in the correlogram non-significative autocorrelations are higher than the bands of significance (+-1.96/sqrt(T)) while in the FAC they aren't.
Cheers
Javi
14 years, 5 months
correlogram
by Javier García
Hi, everybody;
I am trying to adjust an ARMA model to one series but, when I use de correlogram function in Gretl I observe that while the number of significatives lags is 14 at 5%, the correlogram (the graphic) show quite more (25-30).
Cheers
Javi
14 years, 5 months
Fwd: Re: missing pixels for Windows PNG driver
by Alan G Isaac
A suggestion ...
Alan
-------- Original Message --------
Subject: Re: missing pixels for Windows PNG driver
Date: Mon, 02 Nov 2009 11:33:28 -0800
From: sfeam <sfeam(a)users.sourceforge.net>
Reply-To: sfeam(a)users.sourceforge.net
Organization: gnuplot development team
Newsgroups: comp.graphics.apps.gnuplot
We request full hinting for png output but
turn it off for pdf output. You could try this simple change to cairo.trm:
--- ./cairo.trm.buggy 2009-09-13 10:49:16.000000000 -0700
+++ ./cairo.trm 2009-11-02 11:30:02.000000000 -0800
@@ -418,8 +418,8 @@ void cairotrm_init()
surface = cairo_image_surface_create( CAIRO_FORMAT_ARGB32,
plot.device_xmax /*double width_in_points*/,
plot.device_ymax /*double height_in_points*/);
- /* png is bitmapped, let's do the full hinting */
- plot.hinting = 100;
+ /* png should like hinting, but there seems to be a problem */
+ plot.hinting = 0;
/* png is produced by cairo "image" backend, which has full support
* of OPERATOR_SATURATE */
plot.polygons_saturate = TRUE;
14 years, 5 months
unwanted clipping
by Alan G Isaac
I'm seeing clipping of the y-axis label in my graphs.
Known problem?
Thanks,
Alan Isaac
14 years, 5 months
Script & Signs
by Henrique
Dear Gretl comunnity,
I'm estimating and saving a series of OLS models using a
script. Before start to make the regular statistical tests on these
models I need to check the signs of the estimated parameters. It is
something like that: suppose the economic model,
Y = b[1] + b[2]*X +b[3]*Z
The theory tells me that b[2] < 0 and b[3] > 0 and I estimate "n" models,
Model 1: Y = b[1] + b[2]*X1 +b[3]*Z1
Model 2: Y = b[1] + b[2]*X2 +b[3]*Z2
.
.
.
Model n: Y = b[1] + b[2]*Xn +b[3]*Zn
Now I would like to know if it is possible to obtain a list with the
models where the estimated parameters follow the theory (i.e. b[2] < 0
and b[3] > 0).
Best,
Henrique C. de Andrade
Doutorando em Economia Aplicada
Universidade Federal do Rio Grande do Sul
www.ufrgs.br/ppge
14 years, 5 months
