Ok.
I thought you were up for changing something.
There is though one thing there still is a little tricky.
I can't test X, Y values for graphs which are interpolated.
This could for example be for the model DPL94.
They will change "alot".
This is because, that the fitted parameters to the model will change,
for each run.
This is again related to the exponential curves, as then the R2eff
errors are determined by Monte
Carlo simulations which are random.
When there is random errors, the chi-square values will also be
subject to random values,
and hence the minimisation will give rise to random results of the
fitted parameters of the model.
R1rho', phi_ex, kex.
This could be quite confusing for the user.
For every analysis, he/she will get random values and graphs.
Is this a design by purpose?
Methods to solve this, could be:
- Large number of monte carlo simulations, to minimise the difference
in the error between analysis.
- Make it possible to pass a number to the random generator, thereby
being able to "fix" the randomness.
Best
Troels
2014-07-31 9:36 GMT+02:00 Edward d'Auvergne <edward@xxxxxxxxxxxxx>:
Hi Troels,
Why am I assigned to this bug report? And why is it open? If you are
checking the error values in the text of the Grace file, these will
always change. I've run the
Relax_disp.test_r1rho_kjaergaard_auto_check_graphs system test in your
r1rho_plotting branch but the test always passes. The change at
r24864 (http://article.gmane.org/gmane.science.nmr.relax.scm/22614) is
perfect for solving this problem and is exactly the solution I would
have come up with - check that all header lines are identical and,
then for the data lines, check that the X and Y values are identical
and skip the Y_err values. The test is perfect. Therefore I think
this bug can be closed.
Regards,
Edward
On 30 July 2014 23:21, Troels E. Linnet
<NO-REPLY.INVALID-ADDRESS@xxxxxxx> wrote:
URL:
<http://gna.org/bugs/?22409>
Summary: Minimisation of R1rho model give random results
Project: relax
Submitted by: tlinnet
Submitted on: Wed 30 Jul 2014 09:21:29 PM UTC
Category: relax's source code
Specific analysis category: Relaxation dispersion
Priority: 5 - Normal
Severity: 4 - Important
Status: None
Assigned to: bugman
Originator Name:
Originator Email:
Open/Closed: Open
Release: Repository: trunk
Discussion Lock: Any
Operating System: All systems
_______________________________________________________
Details:
This was discussed in:
http://thread.gmane.org/gmane.science.nmr.relax.devel/6538
Systemtest Relax_disp.test_r1rho_kjaergaard_auto_check_graphs
shows that the error value constantly are changing between each
run of analysis.
This is properly because the number of Monte-Carlo simulation is only set
to
3.
This can sometimes happens if the base data are exponential curves, as then
the R2eff errors are determined by Monte Carlo simulations which are
random.
And because it is a test the number of simulations are low so the
randomness
is quite large.
This randomness will then be propagated into the higher models.
Apart from the unique case of Monte Carlo simulations (or the unit tests
for
a
few other functions which use the 'random' Python module), a test
should return the same result to machine precision every single time.
It must be identical.
_______________________________________________________
Reply to this item at:
<http://gna.org/bugs/?22409>
_______________________________________________
Message sent via/by Gna!
http://gna.org/
Re: [bug #22409] Minimisation of R1rho model give random results