Box Plot Identification.
by Luiz Cesar
Dear All,
My name is Luiz Cesar. I´m a new user of Gretl, I have some doubts, that i guess you can help me.
I edited the gnuplot bellow, where can i add the labels?
Thank you very much!!!
Luiz Cesar
# set term pngcairo font "verdana,8"
set encoding utf8
# boxplots
set style line 1 lc rgb "#ff0000"
set style line 2 lc rgb "#0000ff"
set style line 3 lc rgb "#00cc00"
set style line 4 lc rgb "#bf25b2"
set style line 5 lc rgb "#8faab3"
set style line 6 lc rgb "#ffa500"
set style line 8 lc rgb "#dddddd"
set style increment user
set title "CustosOperacion"
set nokey
set xrange [0:2]
set yrange [3.35038e+007:1.93989e+009]
set ytics nomirror
set border 2
set noxtics
set boxwidth 0.3 relative
plot \
'-' using 1:3:2:5:4 w candlesticks lt 2 lw 2 notitle, \
'-' using 1:2:2:2:2 w candlesticks lt -1 notitle, \
'-' using 1:2 w points pt 1 notitle, \
'-' using 1:2 w points lt 0 pt 7 ps 0.25 notitle
1 1.04111e+008 2.09326e+008 4.76164e+008 7.31231e+008
e
1 2.91621e+008
e
1 4.35339e+008
e
# auxdata 3 2
1 1030811332
1 1267776861
1 1869287128
e
13 years
Re: [Gretl-users] AR-GARCH-t MLE
by Daniel Bencik
Thank you all, AR(7) turned out to be the best specification in the end. Trying to fit a short memory model on long memory volatility is a pain.
Allin, do you still want my dataset?
All the best,
Daniel
13 years
matrix and yes no
by clarodina clarodina
How to do a yes no calculation on matrix?
Matrix m has 2 columns class A and class B and 200 rows exam scores 0 - 100
Each row is 1 student
Want to sum the the number of students in class A with scores between 40
and 60
13 years
Re: [Gretl-users] saving data
by Henrique Andrade
Dear Clarodina, please take a look at the script attached. I think now it
solves your problem :-)
Best regards,
Henrique Andrade
2012/3/25 clarodina clarodina <clarodina(a)gmail.com>
> Hi Henrique,
>
> Tks for replying
>
> The data file structure is csv of text file below
>
> The actual file has 6000 rows and 100 columns
>
> The first column is date and second to 100 columns are name d1 to d100
>
> Some columns have blank with no data
>
(...)
Want to do is to use a rolling 10 rows data points or 20 rows data or other
> custom number of rows data points to compute the corr between the d1 - d100
> columns
>
> Compute a ROLLING 20 rows data points of corr(d1,d2) on the 6000 rows of
> d1 d2 data points
>
> Get the output of corr(d1,d2) and determine the number of corr(d1,d2)
> output between 0.2 to 0.6 values
>
> Sum the number of corr(d1,d2) output between 0.2 to 0.6 values and divide
> from the total number of observations or datapoints use on the computation
> of corr
>
> The process is for d1 and d2 columns but would want to extend the
> computation to (d1,d3) (d1,d4) (d1,d5) (d2,d3) (d2,d4) (d2,d5)
>
> The file has some blank rows with no data and gretl may not accept the
> file and some script would require to ignore the blank rows
>
13 years
Re: [Gretl-users] saving data
by clarodina clarodina
>
> Hi Henrique,
>
> Tks for replying
>
> The data file structure is csv of text file below
>
> The actual file has 6000 rows and 100 columns
>
> The first column is date and second to 100 columns are name d1 to d100
>
> Some columns have blank with no data
>
> Date d1 d2 d3 d4 d5 d6 d7 d8 2/8/2011 0.064711
> 0.001324
> 0.000586
>
> 0.001102 2/9/2011 0.065808
> 0.001365
> 0.000667
>
> 0.001101 2/10/2011 0.06506
> 0.0015
> 0.000652
>
> 0.001093 2/11/2011 0.06511
> 0.001492
> 0.000615 0.001886
> 0.001072 2/14/2011 0.063594
> 0.001514
> 0.000629 0.00184
> 0.001113 2/15/2011 0.063345 0.0577 0.00147
> 0.000593 0.001374
> 0.001258 2/16/2011 0.064223 0.05828 0.00147
> 0.000523 0.00142
> 0.001145 2/17/2011 0.062627 0.05525 0.001484 0.000523 0.00041 0.001542
> 0.00159 2/18/2011 0.062208 0.05525 0.001431 0.00041 0.000401 0.001145
> 0.001709 0.001568 2/22/2011 0.061251 0.05502 0.001474 0.000401 0.000452
> 0.00159 0.001925 0.001367 2/23/2011 0.060523 0.05442 0.00148 0.000452
> 0.000502 0.001568 0.001557 0.05458 2/24/2011 0.061141 0.05458 0.001457
> 0.001382 0.000463 0.001367 0.001584 0.05534 2/25/2011 0.062448 0.05534
> 0.001453 0.001075 0.000532 0.001382 0.001217 0.05558 2/28/2011 0.062498
> 0.05558 0.001404 0.001382 0.000431 0.001075 0.001418 0.05515 3/1/2011
> 0.062448 0.05515 0.001384 0.001075 0.000465 0.001101 0.001264 0.05591
> 3/2/2011 0.062837 0.05591 0.001391 0.001285 0.000529 0.001293 0.001152
> 0.064223 3/3/2011 0.064272 0.05655 0.001404 0.001268 0.000565 0.001429
> 0.001101 0.062627 3/4/2011 0.063604 0.05638 0.001374 0.001235 0.00056
> 0.001208 0.000951 0.062208 3/7/2011 0.062866 0.05538 0.001354 0.00125
> 0.00049 0.001153 0.0012 0.061251 3/8/2011 0.063893 0.056 0.001406 0.001244
> 0.000576 0.001613 0.001185 0.060523 3/9/2011 0.064322 0.05596 0.001393
> 0.001205 0.000602 0.00142 0.000938 0.05638
> Want to do is to use a rolling 10 rows data points or 20 rows data or
> other custom number of rows data points to compute the corr between the d1
> - d100 columns
>
> Tks
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
13 years
Re: [Gretl-users] AR-GARCH-t MLE
by Daniel Bencik
Hello everybody,
first off, thank you all for suggestions.
Riccardo, I know that anything bigger than arma(2,1) is going to be a crappy model. Im forecating volatility and arma is one of the options, the obviously sloppy one but nevertheless the benchmark and I can prove nowehere else than on out of sample that its great fit is just an illusion. That's actually the point...
So, the ARMA(6,5)-GARCH(1,1) model with T-distributed residuals in Eviews gives estimates which are to be seen at http://eubie.sweb.cz/gretl_forum/Eviews_estimation.PNG
When I plug in the very same numbers as initial values into my gretl code, I get convergence, however, I arrive at different estimates, see http://eubie.sweb.cz/gretl_forum/gretl_forum_correct_starting_values.txt
When I however change those starting values just a little bit (namely ga, arc and dof by as little as 0.05), convergence is gone, see http://eubie.sweb.cz/gretl_forum/gretl_forum_slightly_incorrect_starting_...
The most interesting point is that Eviews is telling me that LL = 3978 when MLE is found. When I start gretl with those "correct" values, the reported LL = 2036. Hence, I would suspect a difference in LL function. Eviews claims to be using http://www.scribd.com/doc/44246798/EViews-7-Users-Guide-II page 198 Eq 7.11 which I have coded as
mle ll = -0.5*log(pi*(dof-2)) - lngamma(dof/2) + lngamma((dof + 1)/2) -0.5*log(h) - 0.5*(dof + 1)*log(1 + (dof - 2)^(-1)*h^(-1)*e^2)
and I cant see the difference. My script is at http://eubie.sweb.cz/gretl_forum/ARMA65garch11t%20pro%20gretl%20forum.inp
I dont know where to look for a possible error anymore.
Thank you all once again,
Daniel
13 years
Re: [Gretl-users] saving data
by clarodina clarodina
Henrique Andrade,
The script has 20 iterations
The australia.gdt has 77 obs
Using 20 data points for calculation should have > 20 iterations to current
smpl 1986:2 : 1991:1 (n = 20)
but the script has to Current sample: 1977:1 - 1981:4 (n = 20)
want to have the script to use the 20 data points for calculation loop full
obs
tks
13 years
saving data
by clarodina clarodina
sven, currently smpl 1 20 corr smpl+1+1 and corr smpl+1+1 corr on the raw
data for 20 data points
is quite laborious want to is to loop the corr on 20 points and write
the output of corr to a series save on gretl for more other calculation
that why ask what is the accessor to retrive the output of corr and to
write to the series and save
13 years
Re: [Gretl-users] Gretl-users Digest, Vol 62, Issue 20
by clarodina clarodina
did a loop on a number of data points for corr but how to store them on a
series or matrix
anyone has a sample script
On Fri, Mar 23, 2012 at 10:00 AM, <gretl-users-request(a)lists.wfu.edu> wrote:
> 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: AR-GARCH-t MLE (Rodrigo Alfaro Arancibia)
> 2. Re: saving data (Sven Schreiber)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 23 Mar 2012 11:54:46 -0300
> From: Rodrigo Alfaro Arancibia <ralfaro(a)fen.uchile.cl>
> Subject: Re: [Gretl-users] AR-GARCH-t MLE
> To: Gretl list <gretl-users(a)lists.wfu.edu>
> Message-ID:
> <CA+M9tHqYYZ92LvAMAKSv5=PEcg1GEU83LASuAMoBWXVx_jGZNQ(a)mail.gmail.com
> >
> Content-Type: text/plain; charset=ISO-8859-1
>
> 2012/3/23 Allin Cottrell <cottrell(a)wfu.edu>:
> > On Fri, 23 Mar 2012, Riccardo (Jack) Lucchetti wrote:
> >
> >> On Fri, 23 Mar 2012, Daniel Bencik wrote:
> >
> >>> However, Im gettin into a problem (again!). The estimation does not
> >>> converge. For some simpler specifications ( I mean simpler than
> >>> ARMA(7,5)-GARCH(1,1)-t ) I was able to help the algo to converge by
> setting
> >>> different parameter starting values. However, with this complicated
> model,
> >>> even when I set the initial values equal to "true" values found by
> eviews,
> >>> the algo still does not converge.
> >>
> >> With all due respect:
> >>
> >> 1) Unless you have over one billion observations, then estimating an
> >> ARMA(7,5) model is asking for trouble
> >>
> >> 2) I wouldn't put my life in the hands of Eviews' optimising algorithms.
> >
> > To expand on Jack's point 2) a little: it's possible that
> > Eviews' algorithm is more effective in finding the MLE in this
> > case, but it's also possible that Eviews is using a relatively
> > sloppy criterion for reporting a claimed MLE. Before
> > concluding that Eviews is giving you "true" values, you should
> > look at the gradients at "convergence".
> >
> > In gretl, use the --verbose switch with the mle command to see
> > what's happening in detail. You might -- or might not -- see
> > that gretl is finding a log-likelihood comparable to Eviews'
> > maximum, but is not reporting convergence based on the
> > gradient criterion. If you find that Eviews is in fact
> > reaching a proper solution with a greater log-likelihood than
> > gretl is able to find, that would be worth reporting back to
> > us.
> >
> > Allin Cottrell
>
> Daniel, are you sure about 'true-values' in EViews?
> I mean, did the algorithm converge? In other words,
> did it say 'convergence achieved'? If that is the case,
> you could evaluate your LL with zero iteration and
> see the value of gradients (Allin's suggestion).
>
>
> ------------------------------
>
> Message: 2
> Date: Fri, 23 Mar 2012 16:05:48 +0100
> From: Sven Schreiber <svetosch(a)gmx.net>
> Subject: Re: [Gretl-users] saving data
> To: gretl-users(a)lists.wfu.edu
> Message-ID: <4F6C914C.9080409(a)gmx.net>
> Content-Type: text/plain; charset=ISO-8859-1
>
> On 03/23/2012 03:01 PM, Allin Cottrell wrote:
> > On Fri, 23 Mar 2012, clarodina clarodina wrote:
> >
> >> Hi, is there a accessor for correlation?
> >
> > There's the corr() function.
> >
>
> hey, don't give the secret away :-)
>
> -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 62, Issue 20
> *******************************************
>
13 years
