Re: [Gretl-users] on kernel density
Thanks Ricardo,
I have reversed the axes:
<hansl>
matrix d_1 = kdensity(Water)
matrix d[,1] = d_1[,2]
matrix d[,2] = d_1[,1]
<hansl>
Now its ok. I think that the orders in the commmand kdensity are reversed,
or the gnuplot command use the information in the matrix that Gretl create
in reverse form (column 1 in the x axes and column 2 y axes, when it should
be otherwise). Please check the two figures attached.
With this transformation, everything looks fine.
Thanks a lot for the answers and the change in the program
Gretl is the best!
Leandro
2014-01-28 Riccardo (Jack) Lucchetti <r.lucchetti(a)univpm.it>
On Tue, 28 Jan 2014, Leandro Zipitria wrote:
Dear all,
>
> First, I have checked the CSV and confirm that now the kernel works as
> expected.
> Thanks all for the help.
> Just to add that the information is plotted reversed (the data generated
> for the kernel approximation is reversed in the matrix), I attach the file
> generated.
>
Whoops, youre right. Use this:
<hansl>
gnuplot 2 1 --matrix=d --with-lines --output=pruba.eps --suppress-fitted
</hansl>
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------
Attachments:
- attachment.html (text/html — 2.3 KB)
- wrong.eps (application/postscript — 21.1 KB)
- right.eps (application/postscript — 21.1 KB)