There may be a problem with plhist, using v 5.14/gFortran 7.2. If the upper/lower (bin/x-axis) bounds of the hist are too small, then there appear spurious frequencies in the "wings" (i.e. at each of the far ends).
I have tested several distribution. I attach four results for Standard Normal (Gaussian) dist with bounds of (-1,1), (-2,2), (-3,3), and (-5,5). Once the bounds are wide enough, all is well, but otherwise the hist is wrong. The narrower the bounds, the worse the error in the wings.
Those examples are working correctly for the opt=PL_HIST_DEFAULT case (see http://plplot.org/docbook-manual/plplot-html-5.14.0/plhist.html). Therefore, I am going to close this support request. However, when I systematically looked at other opt values for a modified version of C example 5, I found there were some bugs in our plhist C library code which I will try to sort out today. So watch our git log for those fixes. And those fixes should automatically propagate to our Fortran binding for that library if you use the git master tip version of PLplot (or 5.15.0 when it comes out).
No, actually it's not working correctly for the default example 5 either.
The issue there is that the "oscillator's" correct result is a "convex"
profile, and the "error is in the convexity" so to speak, so it is more
difficult to see the error.
That is why I used a Gaussian example to illustrate the problem, since it is
immediately obvious something is wrong.
I attach a sequence of results using the original example 5 code, for a
range of boundaries from the original (-1.1,1,1) up to (-5.1,5.1), attached
as jpgs. Notice the error/misbehaviour in the at-the-boundary, and 1-in
from the boundary bars.
Cheers
DrO
-----Original Message-----
From: Alan W. Irwin [mailto:airwin@users.sourceforge.net]
Sent: 21 December 2018 16:57
To: [plplot:support-requests]
Subject: [plplot:support-requests] #46 plhist seems odd v 5.14
Those examples are working correctly for the opt=PL_HIST_DEFAULT case (see
http://plplot.org/docbook-manual/plplot-html-5.14.0/plhist.html). Therefore,
I am going to close this support request. However, when I systematically
looked at other opt values for a modified version of C example 5, I found
there were some bugs in our plhist C library code which I will try to sort
out today. So watch our git log for those fixes. And those fixes should
automatically propagate to our Fortran binding for that library if you use
the git master tip version of PLplot (or 5.15.0 when it comes out).
[support-requests:#46]
https://sourceforge.net/p/plplot/support-requests/46/ plhist seems odd v
5.14
Status: open
Group:
Created: Fri Dec 21, 2018 03:40 PM UTC by DrO
Last Updated: Fri Dec 21, 2018 03:40 PM UTC
Owner: nobody
Attachments:
x0f5_Hist_1.jpg
https://sourceforge.net/p/plplot/support-requests/46/attachment/x0f5_Hist_1 .jpg (41.0 kB; image/jpeg)
x0f5_Hist_2.jpg
https://sourceforge.net/p/plplot/support-requests/46/attachment/x0f5_Hist_2 .jpg (58.4 kB; image/jpeg)
x0f5_Hist_3.jpg
https://sourceforge.net/p/plplot/support-requests/46/attachment/x0f5_Hist_3 .jpg (52.1 kB; image/jpeg)
x0f5_Hist_5.jpg
https://sourceforge.net/p/plplot/support-requests/46/attachment/x0f5_Hist_5 .jpg (45.1 kB; image/jpeg)
There may be a problem with plhist, using v 5.14/gFortran 7.2. If the
upper/lower (bin/x-axis) bounds of the hist are too small, then there appear
spurious frequencies in the "wings" (i.e. at each of the far ends).
I have tested several distribution. I attach four results for Standard
Normal (Gaussian) dist with bounds of (-1,1), (-2,2), (-3,3), and (-5,5).
Once the bounds are wide enough, all is well, but otherwise the hist is
wrong. The narrower the bounds, the worse the error in the wings.
Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/plplot/support-requests/46/
To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/
Related
Support Requests:
#46