Discussion:
Lograithmic plotting of ccdf
(too old to reply)
Detlef Bosau
2008-01-07 11:02:31 UTC
Permalink
Hi.

I want to create a logarithmic plot of a ccdf.
My octave version 2.1.73


Because I´m primarily interested in the small probabilities, I would
like to have a plot as I have seen it in a paper: The y-axis was not
scaled from 0 to 1 but logarithmic from, say, -4 to 0 which refers to
probabilities 10^-4 to 10^0 = 1.

Refer to http://www.detlef-bosau.de/eurane_results/delay_s_1_t_0.1.eps.gz
as an example for what I want to do.

Can someone please help me on this one?

Thanks.

Detlef
--
Horst Nietowski: Ingrid ist doof.
Bernd Nawothnig: Sie ist bosau. Schade, dass dieser Liebe(ler) keine
Zukunft gegönnt ist.
25. Okt. 2007, <***@bernd.nawothnig.dialin.t-online.de>
Francesco Potorti`
2008-01-07 16:18:25 UTC
Permalink
Post by Detlef Bosau
Because I´m primarily interested in the small probabilities, I would
like to have a plot as I have seen it in a paper: The y-axis was not
scaled from 0 to 1 but logarithmic from, say, -4 to 0 which refers to
probabilities 10^-4 to 10^0 = 1.
At the Octave prompt, type
help semilogy
Detlef Bosau
2008-01-17 21:07:36 UTC
Permalink
Post by Francesco Potorti`
Post by Detlef Bosau
Because I´m primarily interested in the small probabilities, I would
like to have a plot as I have seen it in a paper: The y-axis was not
scaled from 0 to 1 but logarithmic from, say, -4 to 0 which refers to
probabilities 10^-4 to 10^0 = 1.
At the Octave prompt, type
help semilogy
May I ask an additional question?

I want to use lines and points as well, IIRC there is some style
"linespoints" in gnuplot, but I don´t want to plot every point with
"points" but perhpas every 10th. (IIRC, there is some keyword "every" in
Gnuplot exactly for this purpose.)


So, in plots like this:
Loading Image...

or as postscript:
http://www.detlef-bosau.de/eurane_results/delay_s_2_t_0.1.eps.gz

I want to distinguish the plots e.g. by circles and triangles or
something like that, but only every 10th point should be plotted as,
e.g., circle.

Please allow me to put my plot file here:

============================8<======================================


load delay0;
load delay1;
load delay2;
load delay3;
load delay4;
load delay5;
load delay6;
load delay7;
x=0;
for i = 1:500;
x = [x, i/10];
endfor;
xlabel("TCP Packet Delay [s]")
ylabel("CCDF")
gset title "TCP Packet Delay, Scheduler: Elastic Round Robin, Tolerance:
0.01"
gset term postscript color
gset output "delay_s_11_t_0.01.eps"
gset yrange [0.001:1]
semilogy(x,1-empirical_cdf(x,delay0),"-@11;delay 0;"
,x,1-empirical_cdf(x,delay1),"-@21;delay 1;"
,x,1-empirical_cdf(x,delay2),"-@31;delay 2;"
,x,1-empirical_cdf(x,delay3),"-@41;delay 3;"
,x,1-empirical_cdf(x,delay4),"-@13.;delay 4;"
,x,1-empirical_cdf(x,delay5),"-@23;delay 5;"
,x,1-empirical_cdf(x,delay6),"-@33;delay 6;"
,x,1-empirical_cdf(x,delay7),"-@43;delay 7;");



===========================8<=======================================

Is there an easy way to achieve this?

Many thanks!

Detlef
--
Horst Nietowski: Ingrid ist doof.
Bernd Nawothnig: Sie ist bosau. Schade, dass dieser Liebe(ler) keine
Zukunft gegönnt ist.
25. Okt. 2007, <***@bernd.nawothnig.dialin.t-online.de>
Loading...