removing nan and inf from a matrix
by Logan Kelly
Hello,
I need to take the log difference of a matrix, i.e. log(M[2 rows(M):,]/M[1:rows(M)-1,]). Unfortunately, M has elements equal to zero. I need to replace the nan's and inf's with 0's. This almost works
M = isnan(M) ? 0 : M
but does not remove inf's. Any sugestions?
9 years
quantile() definition details
by Sven Schreiber
Hi,
I wonder whether the behavior of the quantile() function in cases where
p < 1/(n+1) is optimal.
Currently, things like quantile({-1; 1}, 0.1) return NA, because 0.1 <
rows({-1; 1}). (See the definition in the quantile reference, here
(n+1)*p = 3*0.1 = 0.3, of which the integer part is zero, and there
exists no zero-th element in the ordered data vector.)
Given that quantiles can be defined in different ways, this is not
wrong. However, one could also use the minimum value in these cases.
This would correspond to a quantile definition where a fraction p of the
values is smaller *or equal to* the relevant quantile.
See also the practice in other software, for example the repeated phrase
"When p < 1 / (N+1), use x1" in the following software survey table:
http://en.wikipedia.org/wiki/Quantile#Estimating_the_quantiles_of_a_popul...
So my bottom line is: quantile() should always return something useful
instead of NA.
Thanks,
sven
10 years
Feature Request: "--preserve" option
by Henrique Andrade
Dear Gretl Team,
We can use the option "--preserve" with the "open" command to preserve
matrices and scalars. I would like to ask you if it is possible to include
strings too (so the "preserve" option will preserve matrices, scalars, and
strings).
Best regards,
Henrique Andrade
10 years
test-down option reports different results in adf and coint
by Koray Simsek
Dear All,
I've been trying to figure out why the test-down option behaves differently
in coint and adf. In the help file, it is explained that the approach in
adf is used in coint as well, but the output is inconsistent with this for
the initial DF steps in coint (steps 1 and 2).
I would expect both functions to run the adf on a subset of the data (for
comparability reasons, I presume) that is consistent with the maximum lag.
Initially, this appears to be the case as the MAIC values for both adf and
coint are identical. BUT after finding the number of lags with minimum
MAIC, adf reports the results for the full set wheres coint reports the
results for the subset of data where original estimation is done.
I noticed a few threads in which the logic of --test-down was discussed
(whether the maximum possible data is used for each lag vs the current
set-up) but could not really find anything with respect to this particular
difference in adf and coint. I am not sure whether this is a bug or not.
If you would like to replicate this situation, please load denmark.gdt and
then run
adf 5 LRM --c --test-down --verbose
coint 5 LRM LRY --test-down --verbose
Although the MAIC values are the same in both calls, adf reports the ADF
regression on T=53 and coint reports the ADF regression on T=49.
Which output refers to the one with the minimum MAIC? Is this a bug or is
there a specific reason behind this choice? My gretl version is 1.9.92.
Best,
Koray
10 years
Re: [Gretl-users] Gretl-users Digest, Vol 100, Issue 22
by Liliya Netskar
Hello Sven,
thank you very much for your answer.
I see this. But I still don't know - taking this example - how I can control for individual fixed effects for Joe, Sue, Mark with Gretl?
Liliya
"gretl-users-request(a)lists.wfu.edu" <gretl-users-request(a)lists.wfu.edu> schrieb am 18:01 Dienstag, 26.Mai 2015:
Send Gretl-users mailing list submissions to
gretl-users(a)lists.wfu.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.wfu.edu/mailman/listinfo/gretl-users
or, via email, send a message with subject or body 'help' to
gretl-users-request(a)lists.wfu.edu
You can reach the person managing the list at
gretl-users-owner(a)lists.wfu.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Gretl-users digest..."
Today's Topics:
1. Re: Gretl-users Digest, Vol 100, Issue 20 (Sven Schreiber)
----------------------------------------------------------------------
Message: 1
Date: Tue, 26 May 2015 01:53:54 +0200
From: Sven Schreiber <svetosch(a)gmx.net>
To: gretl-users(a)lists.wfu.edu
Subject: Re: [Gretl-users] Gretl-users Digest, Vol 100, Issue 20
Message-ID: <5563B612.3000602(a)gmx.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Am 25.05.2015 um 21:13 schrieb Liliya Netskar:
> Actually I used a very simple data set for the purporse of learning. The
> Gretl output file and the data set are attached below. Sorry, I have a
> german version :). Thanks again for your help!
>
Your output (parts):
Modell 1: Fixed-effects, mit 12 Beobachtungen
Mit 4 Querschnittseinheiten
Zeitreihenl????nge = 3
Abh????ngige Variable: TestGrade
Weggelassen wg. exakter Kollinearit????t: Joe Sue Mark
So if I interpret your setup correctly, you told gretl to include unit
fixed effects (hence gretl's description as "Fixed-effects"), and then
you manually created and included unit dummies "Joe", "Sue", and "Mark".
So effectively you're throwing in two sets of identical (or collinear)
unit dummies. Of course they're collinear then and gretl throws them out
because of that.
(The fact that the calculation is done with the within transformation
doesn't matter for this problem.)
hth,
sven
------------------------------
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
End of Gretl-users Digest, Vol 100, Issue 22
********************************************
10 years, 1 month
Re: [Gretl-users] Gretl-users Digest, Vol 100, Issue 20
by Liliya Netskar
Actually I used a very simple data set for the purporse of learning. The Gretl output file and the data set are attached below. Sorry, I have a german version :). Thanks again for your help!
Liliya
"gretl-users-request(a)lists.wfu.edu" <gretl-users-request(a)lists.wfu.edu> schrieb am 18:02 Montag, 25.Mai 2015:
Send Gretl-users mailing list submissions to
gretl-users(a)lists.wfu.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://lists.wfu.edu/mailman/listinfo/gretl-users
or, via email, send a message with subject or body 'help' to
gretl-users-request(a)lists.wfu.edu
You can reach the person managing the list at
gretl-users-owner(a)lists.wfu.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Gretl-users digest..."
Today's Topics:
1. Fixed effects model (Liliya Netskar)
2. Re: Fixed effects model (Sven Schreiber)
----------------------------------------------------------------------
Message: 1
Date: Mon, 25 May 2015 11:47:32 +0000 (UTC)
From: Liliya Netskar <liliya_netskar(a)yahoo.com>
To: "gretl-users(a)lists.wfu.edu" <gretl-users(a)lists.wfu.edu>
Subject: [Gretl-users] Fixed effects model
Message-ID:
<1847116116.2539701.1432554452728.JavaMail.yahoo(a)mail.yahoo.com>
Content-Type: text/plain; charset="utf-8"
Hi guys!
I have a problem with Gretl panel fixed effects model. I created dummy variables to control for fixed effects (it was a proper way), but Gretl leaves these dummies out. To represent 4 categoeries I used 3 dummies to avoid perfect multicolinearity. In the end I habe a biased result because of the omitted dummies. Is there a solution for the problem?
Thanks a lot in advance!
Liliya
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wfu.edu/pipermail/gretl-users/attachments/20150525/50c9591f/...>
------------------------------
Message: 2
Date: Mon, 25 May 2015 14:17:43 +0200
From: Sven Schreiber <svetosch(a)gmx.net>
To: gretl-users(a)lists.wfu.edu
Subject: Re: [Gretl-users] Fixed effects model
Message-ID: <556312E7.7080101(a)gmx.net>
Content-Type: text/plain; charset=windows-1252; format=flowed
Am 25.05.2015 um 13:47 schrieb Liliya Netskar:
> Hi guys!
>
> I have a problem with Gretl panel fixed effects model. I created dummy
> variables to control for fixed effects (it was a proper way), but Gretl
> leaves these dummies out. To represent 4 categoeries I used 3 dummies to
> avoid perfect multicolinearity. In the end I habe a biased result
> because of the omitted dummies. Is there a solution for the problem?
>
Gretl typically leaves out regressors (including dummies) for a reason
-- usually collinearity. In the case of panel fixed effects one
immediately thinks of time-invariant regressors as a reason, but this is
of course just speculation without knowing any details. Please show us
the estimation output.
cheers,
sven
------------------------------
_______________________________________________
Gretl-users mailing list
Gretl-users(a)lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users
End of Gretl-users Digest, Vol 100, Issue 20
********************************************
10 years, 1 month
Fixed effects model
by Liliya Netskar
Hi guys!
I have a problem with Gretl panel fixed effects model. I created dummy variables to control for fixed effects (it was a proper way), but Gretl leaves these dummies out. To represent 4 categoeries I used 3 dummies to avoid perfect multicolinearity. In the end I habe a biased result because of the omitted dummies. Is there a solution for the problem?
Thanks a lot in advance!
Liliya
10 years, 1 month
MODIFIED AIC
by Dipankar Basnet
HI,
I AM USING GRETL FOR MAC (QUARTZ). WHEN I CONDUCT THE ADF TEST, THE TEST DOWN FOR LAG ORDER DOES NOT DISPLAY THE “MODIFIED” AIC AND BIC RATHER AIC AND BIC. THE RESULTS WHEN I COMPARE ARE DIFFERENT FOR MAC (WITHOUT “MODIFIED” AIC) THAN THE MODIFIED AIC THAT IS AVAILABLE WHEN USING GRETL FOR WINDOWS. I WOULD BE GRATEFUL IF YOU COULD SUGGEST A SOLUTION TO THIS CONFUSION.
THANKS,
DIPANKAR
10 years, 1 month
plotting time-series with two different scales
by ziggy s
Hi,
I'm fairly new to gretl. It's a good tool that helps me with my work.
I have a question about plotting time series (daily data) with the date as
labels on the x axis.
I've noticed that when using the --time-seires option in gnuplot, the plot
is always against a single data scale (on the left).
In other words, the option --time-series, seems to imply --single-yaxis.
This behavior is not documented in the User's Guide and I want to ask if
there is a way to undo this behavior and let a gnuplot with --time-series
use both left and right axes. If I don't use --time-series, I can still
plot versus an index column but then the x-labels are not dates.
Here's an example to illustrate:
open anscombe.gdt
setobs 7 2015-02-18 --time-series # assign some arbitrary date range to
sample data
genr index
y100=y1*100 # scale the data by 100 to get different value ranges
gnuplot x y100 --time-series --with-lp # both series on left axis scale
gnuplot x y100 index --with-lp # second series scale is separate (Right
axis)
Another note, I had an idea that a potential workaround could be adding a
unix-timestamp column to my data (I can easily do that with my data source)
and then telling gnuplot to format the x axis as a date. Something like:
gnuplot y1 y2 unix_day --with-lines {gnuplot commands;}
where unix_day is the unix-timestamp column.
Can anyone with gnuplot experience tell me what commands need to go in the
braces for this to work?
I'm using gretl 1.10.1 on Windows 7.
10 years, 1 month
plotting time-series with two different scales
by shahar siegman
Hi,
I'm fairly new to gretl. It's a good tool that helps me with my work.
I have a question about plotting time series (daily data) with the date as
labels on the x axis.
I've noticed that when using the --time-seires option in gnuplot, the plot
is always against a single data scale (on the left).
In other words, the option --time-series, seems to imply --single-yaxis.
This behavior is not documented in the User's Guide and I want to ask if
there is a way to undo this behavior and let a gnuplot with --time-series
use both left and right axes. If I don't use --time-series, I can still
plot versus an index column but then the x-labels are not dates.
Here's an example to illustrate:
open anscombe.gdt
setobs 7 2015-02-18 --time-series # assign some arbitrary date range to
sample data
genr index
y100=y1*100 # scale the data by 100 to get different value ranges
gnuplot x y100 --time-series --with-lp # both series on left axis scale
gnuplot x y100 index --with-lp # second series scale is separate (Right
axis)
Another note, I had an idea that a potential workaround could be adding a
unix-timestamp column to my data (I can easily do that with my data source)
and then telling gnuplot to format the x axis as a date. Something like:
gnuplot y1 y2 unix_day --with-lines {gnuplot commands;}
where unix_day is the unix-timestamp column.
Can anyone with gnuplot experience tell me what commands need to go in the
braces for this to work?
I'm using gretl 1.10.1 on Windows 7.
10 years, 1 month
