3:01 p.m.
It is possible to save an equation version in the quantile regression
option of Robust methods modelling. However, there is nothing in the User
guide to indicate which quantile it relates to.
Any advice welcome
thanks
Brian
Brian J Revell
Professor Emeritus
Agricultural and Food Economics
Harper Adams University
Newport
Shropshire TF10 8NB
Tel 01952 815237
Tel: +44 1952 728153
Mbl +44 7976 538712
University: +44 1952 815235
alt: email: bjrevell(a)harper-adams.ac.uk
3:39 p.m.
Am 11.01.2021 um 16:01 schrieb Brian Revell:
It is possible to save an equation version in the quantile
regression
option of Robust methods modelling. However, there is nothing in the
User guide to indicate which quantile it relates to.
Well, the equation view is really barebones, it doesn't even mention
what kind of model it is in the first place. In my view even the
information there about the residual (abs.) sum could be removed, don't
know why that one was chosen to be shown.
But the quantile is chosen by the user at the specification stage.
Default is the median (0.5). This is shown as tau in the standard model
output.
cheers
sven
4:15 p.m.
Hi
it is not the 0,5 quantile equation as that has been one estimated. The
result seems to lay between 0.1 and 0.2 quantiles
On another matter - Basic Structural Model - error message saying the
forecast function will not work with exogenous variables (even though
values have been supplied for the forecast horizon)
Second error -editing graphs in the ARIMA modelling option - the graph
options for fonts, titles, colours, line styles can all be selected.
However on issuing the Apply command -the following error message is given
"C:\Program Files\gretl\wgnuplot.exe"
"C:\Users\Brian\AppData\Roaming\gretl\gpttmp.7XAUW0": exit code 1
Brian
On Mon, 11 Jan 2021 at 15:39, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 11.01.2021 um 16:01 schrieb Brian Revell:
> It is possible to save an equation version in the quantile regression
> option of Robust methods modelling. However, there is nothing in the
> User guide to indicate which quantile it relates to.
Well, the equation view is really barebones, it doesn't even mention
what kind of model it is in the first place. In my view even the
information there about the residual (abs.) sum could be removed, don't
know why that one was chosen to be shown.
But the quantile is chosen by the user at the specification stage.
Default is the median (0.5). This is shown as tau in the standard model
output.
cheers
sven
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
4:37 p.m.
I estimated 0.1 0.2 0.5 and 0.9 quantiles. THe equation constant and slope
appeared to fall between the 0.1 and 0.2 quantiles. I first compared the
equation coefficients with the 0.5 quantile estimates-but there was no
similarity with the coefficients
Hope that is clear.
On Mon, 11 Jan 2021 at 16:30, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 11.01.2021 um 17:15 schrieb Brian Revell:
> Hi
> it is not the 0,5 quantile equation as that has been one estimated. The
> result seems to lay between 0.1 and 0.2 quantiles
Sorry, I don't understand what you mean.
-s
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
4:56 p.m.
Am 11.01.2021 um 17:37 schrieb Brian Revell:
I estimated 0.1 0.2 0.5 and 0.9 quantiles. THe equation constant and
slope appeared to fall between the 0.1 and 0.2 quantiles. I first
compared the equation coefficients with the 0.5 quantile estimates-but
there was no similarity with the coefficients
Hope that is clear.
OK, when I estimate several quantiles at once (didn't remember that that
was possible!) I can confirm that when I choose (from the model window)
File -> View as equation there is a _single_ equation shown which has
mysterious coefficients, meaning they do not or only partly match
coefficients in the full output.
Could be a bug alright, or else I agree it needs to be explained! Thanks
for reporting it.
cheers
sven
7:18 p.m.
On Mon, 11 Jan 2021, Sven Schreiber wrote:
Am 11.01.2021 um 17:37 schrieb Brian Revell:
> I estimated 0.1 0.2 0.5 and 0.9 quantiles. THe equation constant and
> slope appeared to fall between the 0.1 and 0.2 quantiles. I first
> compared the equation coefficients with the 0.5 quantile estimates-but
> there was no similarity with the coefficients
> Hope that is clear.
OK, when I estimate several quantiles at once (didn't remember that that
was possible!) I can confirm that when I choose (from the model window)
File -> View as equation there is a _single_ equation shown which has
mysterious coefficients, meaning they do not or only partly match
coefficients in the full output.
Could be a bug alright, or else I agree it needs to be explained! Thanks
for reporting it.
"View as equation" should not be enabled for multi-tau quantreg
estimates. That's now fixed in git. What was shown was actually an
initial set of OLS estimates, which are just for internal use.
Allin
7:36 p.m.
Thanks for clarification . It was simply curiosity that made me look at the
equation option.!
On Mon, 11 Jan 2021 at 19:18, Allin Cottrell <cottrell(a)wfu.edu> wrote:
On Mon, 11 Jan 2021, Sven Schreiber wrote:
> Am 11.01.2021 um 17:37 schrieb Brian Revell:
>> I estimated 0.1 0.2 0.5 and 0.9 quantiles. THe equation constant and
>> slope appeared to fall between the 0.1 and 0.2 quantiles. I first
>> compared the equation coefficients with the 0.5 quantile estimates-but
>> there was no similarity with the coefficients
>> Hope that is clear.
>
> OK, when I estimate several quantiles at once (didn't remember that that
> was possible!) I can confirm that when I choose (from the model window)
> File -> View as equation there is a _single_ equation shown which has
> mysterious coefficients, meaning they do not or only partly match
> coefficients in the full output.
>
> Could be a bug alright, or else I agree it needs to be explained! Thanks
> for reporting it.
"View as equation" should not be enabled for multi-tau quantreg
estimates. That's now fixed in git. What was shown was actually an
initial set of OLS estimates, which are just for internal use.
Allin_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
4:31 p.m.
New subject: StrucTiSM (was Re: Re: A query ...)
Am 11.01.2021 um 17:15 schrieb Brian Revell:
On another matter - Basic Structural Model - error message saying
the
forecast function will not work with exogenous variables (even though
values have been supplied for the forecast horizon)
[trying to disentangle the various topics here]
Yes, known missing feature, as documented.
-s
4:35 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
Am 11.01.2021 um 17:15 schrieb Brian Revell:
Second error -editing graphs in the ARIMA modelling option - the
graph
options for fonts, titles, colours, line styles can all be selected.
However on issuing the Apply command -the following error message is given
"C:\Program Files\gretl\wgnuplot.exe"
"C:\Users\Brian\AppData\Roaming\gretl\gpttmp.7XAUW0": exit code 1
cannot replicate this here, need more detailed information about the
setup (versions...) and exact steps
-s
4:46 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
Version: "You are running gretl 2020e. This is the current version."
The error appears when I want to add a line onto the graph -the option is
provided -it asks for the function ( say Z = 12) . Then it gives me the
error message when I issue the Apply command.
B
On Mon, 11 Jan 2021 at 16:35, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 11.01.2021 um 17:15 schrieb Brian Revell:
> Second error -editing graphs in the ARIMA modelling option - the graph
> options for fonts, titles, colours, line styles can all be selected.
> However on issuing the Apply command -the following error message is
given
> "C:\Program Files\gretl\wgnuplot.exe"
> "C:\Users\Brian\AppData\Roaming\gretl\gpttmp.7XAUW0": exit code 1
cannot replicate this here, need more detailed information about the
setup (versions...) and exact steps
-s
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
4:54 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
To follow up -the error relates to trying to add an addition line which is
actually a constant value of the dependent variable . I have tried y -
12.8, SEDE_UNADJ =12.8 which is the depvar or CL =12.8 whiich is also one
of the variables in the Gretl variable list,.. Each time error message
appears. It would seem the option graph actual and fitted against time is
not set up to accept additional variables.
On Mon, 11 Jan 2021 at 16:35, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 11.01.2021 um 17:15 schrieb Brian Revell:
> Second error -editing graphs in the ARIMA modelling option - the graph
> options for fonts, titles, colours, line styles can all be selected.
> However on issuing the Apply command -the following error message is
given
> "C:\Program Files\gretl\wgnuplot.exe"
> "C:\Users\Brian\AppData\Roaming\gretl\gpttmp.7XAUW0": exit code 1
cannot replicate this here, need more detailed information about the
setup (versions...) and exact steps
-s
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
4:59 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
Am 11.01.2021 um 17:54 schrieb Brian Revell:
To follow up -the error relates to trying to add an addition line
which
is actually a constant value of the dependent variable . I have tried y
- 12.8, SEDE_UNADJ =12.8 which is the depvar or CL =12.8 whiich is also
one of the variables in the Gretl variable list,.. Each time error
message appears. It would seem the option graph actual and fitted
against time is not set up to accept additional variables.
I seem to vaguely remember some recent problems there that might be
related. In this case it would be useful if you could check the behavior
of the release candidate a.k.a. the latest snapshot. If you're on
Windows that would (most likely) be this thing:
http://sourceforge.net/projects/gretl/files/snapshots/gretl_install-64.ex...
thanks
sven
6:17 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
OK will do. Not sure the current version is completely stable as it just
cut out of my latest session when I attempted to modify a variable, I have
had several other instances of crash with no warning too so everything
saved as icons are alost as well.
B
On Mon, 11 Jan 2021 at 17:00, Sven Schreiber <svetosch(a)gmx.net> wrote:
Am 11.01.2021 um 17:54 schrieb Brian Revell:
> To follow up -the error relates to trying to add an addition line which
> is actually a constant value of the dependent variable . I have tried y
> - 12.8, SEDE_UNADJ =12.8 which is the depvar or CL =12.8 whiich is also
> one of the variables in the Gretl variable list,.. Each time error
> message appears. It would seem the option graph actual and fitted
> against time is not set up to accept additional variables.
>
I seem to vaguely remember some recent problems there that might be
related. In this case it would be useful if you could check the behavior
of the release candidate a.k.a. the latest snapshot. If you're on
Windows that would (most likely) be this thing:
http://sourceforge.net/projects/gretl/files/snapshots/gretl_install-64.ex...
thanks
sven
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
10:09 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
Am 11.01.2021 um 19:17 schrieb Brian Revell:
OK will do. Not sure the current version is completely stable as it
just
cut out of my latest session when I attempted to modify a variable, I
have had several other instances of crash with no warning too so
everything saved as icons are alost as well.
I'd say this is certainly not normal, as many people are using gretl
regularly, and such behavior would surely drive them (and us) crazy.
That doesn't mean I don't believe you, and if I remember correctly one
of the last times you pointed out some other problem it identified a bug.
So it would be helpful again if you could describe as precisely as
possible what exactly you are doing and what happens.
thanks
sven
8:48 p.m.
New subject: About Pooled Mean Group estimation
Am 12.01.2021 um 18:26 schrieb Reynaldo Senra:
PMG_estimate(&b, , {0;0;0},1000)
Most of the time, I achieve the exact results of Eviews by setting {1;0;0}.
Unfortunately, in one case it has been impossible and I really tried
several combinations. Here I just attached a csv file with the data
involved in the problem. I also attached the Eviews output and the Gretl
sample script.
Hi Reynaldo, that's an interesting case study for the PMG package. First
let me note that the automatic setting for the initial values yields a
better likelihood than either {0,0,0} or {1,0,0} in this case.
Anyway, if you try:
PMG_estimate(&b, , {0.5,0.5,-1}, 1000)
then you will get the same results as Eviews. But that's of course
cheating, because I picked those values in the neighborhood of the
results I saw. So that means that in general we're not clever enough
choosing our initial values, and here we apparently got stuck in a local
maximum.
I'm also noticing, however, that the parameters of the mean-group
estimator {0.89, -0.46, -1.8} lead us to the (presumably) correct PMG
result again. Maybe we should routinely also try MG as an
initialization. Jack?
cheers
sven
9:29 p.m.
New subject: ARIMA plots (was Re: Re: A query ...)
On Mon, 11 Jan 2021, Brian Revell wrote:
To follow up -the error relates to trying to add an addition line
which is
actually a constant value of the dependent variable . I have tried y -
12.8, SEDE_UNADJ =12.8 which is the depvar or CL =12.8 whiich is also one
of the variables in the Gretl variable list,.. Each time error message
appears.
Brian, if you want to add a horizontal line, just enter the
numerical value into the formula box, e.g.
12.8
Basically the only other thing you can do in that box is type a
formula that references "x" (a hard-wired "x", meaning the x-axis),
as in
sin(x)
"y" or any other identifiers are unknown to gnuplot and will cause
an error.
Allin
11:15 a.m.
New subject: ARIMA plots (was Re: Re: A query ...)
Thanks -it worked . it was the only alternative I hadn't tried!
On Mon, 11 Jan 2021 at 21:29, Allin Cottrell <cottrell(a)wfu.edu> wrote:
On Mon, 11 Jan 2021, Brian Revell wrote:
> To follow up -the error relates to trying to add an addition line which
is
> actually a constant value of the dependent variable . I have tried y -
> 12.8, SEDE_UNADJ =12.8 which is the depvar or CL =12.8 whiich is also one
> of the variables in the Gretl variable list,.. Each time error message
> appears.
Brian, if you want to add a horizontal line, just enter the
numerical value into the formula box, e.g.
12.8
Basically the only other thing you can do in that box is type a
formula that references "x" (a hard-wired "x", meaning the x-axis),
as in
sin(x)
"y" or any other identifiers are unknown to gnuplot and will cause
an error.
Allin
_______________________________________________
Gretl-users mailing list -- gretl-users(a)gretlml.univpm.it
To unsubscribe send an email to gretl-users-leave(a)gretlml.univpm.it
Website:
https://gretlml.univpm.it/postorius/lists/gretl-users.gretlml.univpm.it/
1199
days inactive
1200
days old
18 comments
4 participants
participants (4)
-
Allin Cottrell -
Brian Revell -
Reynaldo Senra -
Sven Schreiber