Hi,
Would you be able to print the AIC and k values as well? k is the
number of parameters in the model. The places where the chi-squared
value increases rather than decreases is because of a collapse in
model complexity. If you plot the chi2, AIC, and k values verses
iteration number, like I did in my thesis in figures 7.3 and 7.4
(http://eprints.infodiv.unimelb.edu.au/archive/00002799/), you'll see
what is happening there. The plots should help in figuring out
exactly what is happening.
Regards,
Edward
On 6/25/07, Douglas Kojetin <[EMAIL PROTECTED]> wrote:
> Update: I figured out the problem. I needed to use the 'opt'
> directory for the results.read() call, not 'aic':
>
> #--start of GOOD code
> import glob
> runs = glob.glob('prolate/round_*')
>
> # Loop over the runs.
> for name in runs:
> print name
> run.create(name, 'mf')
> results.read(run=name, file='results', dir=name+'/opt')
> chisq=self.relax.data.chi2[name]
> print "%s: %1.10f" % (name, chisq)
> #--end
>
> Here is the list of chi2 values. I used 30 decimal points for the
> chi2 values -- let me know if you would like something different.
>
> """
> prolate/round_1: 785.330531871414336819725576788187
> prolate/round_2: 786.656854782415166482678614556789
> prolate/round_3: 784.104495289329975094005931168795
> prolate/round_4: 783.543316702498373160779010504484
> prolate/round_5: 786.500523476859029869956430047750
> prolate/round_6: 784.433290432082458210061304271221
> prolate/round_7: 786.264734828735640803643036633730
> prolate/round_8: 785.887140331052023611846379935741
> prolate/round_9: 785.887140331170371609914582222700
> prolate/round_10: 785.887140331282466831908095628023
> prolate/round_11: 785.887140331283262639772146940231
> prolate/round_12: 785.887140331282807892421260476112
> prolate/round_13: 785.887140331283376326609868556261
> prolate/round_14: 785.887140331282921579258982092142
> prolate/round_15: 785.887140331282353145070374011993
> prolate/round_16: 785.887140331283262639772146940231
> prolate/round_17: 785.887140331052364672359544783831
> prolate/round_18: 785.887140331284172134473919868469
> prolate/round_19: 785.887140331283262639772146940231
> prolate/round_20: 785.887140331282694205583538860083
> prolate/round_21: 785.887140331284967942337971180677
> prolate/round_22: 785.887140331337491261365357786417
> prolate/round_23: 785.887140331283944760798476636410
> prolate/round_24: 785.887140331283376326609868556261
> prolate/round_25: 785.887140331282921579258982092142
> prolate/round_26: 785.887140331282353145070374011993
> prolate/round_27: 785.887140331283262639772146940231
> prolate/round_28: 785.887140331052364672359544783831
> prolate/round_29: 785.887140331284172134473919868469
> prolate/round_30: 785.887140331283262639772146940231
> prolate/round_31: 785.887140331282694205583538860083
> """
>
>
> On Jun 25, 2007, at 3:10 AM, Edward d'Auvergne wrote:
>
> > On 6/25/07, Douglas Kojetin <[EMAIL PROTECTED]> wrote:
> >> Hi Edward,
> >>
> >> I'm trying to write a script to calculate the chisq values for each
> >> of the prolate rounds, but I'm having some trouble as I'm not 100%
> >> familiar with the relax subroutines.
> >
> > Your code is almost there. I think I know where the script failed
> > though.
> >
> >
> >> #--start
> >> import glob
> >> runs = glob.glob('prolate/round_*')
> >>
> >> # Loop over the runs.
> >> for name in runs:
> >> name=name+'/aic'
> >> run.create(name, 'mf')
> >> results.read(name)
> >
> > This should all be ok.
> >
> >> chi2=self.relax.data.chi2[run]
> >
> > Here the 'run' variable should be replaced by 'name'.
> >
> >
> >> print "%s: %1.20f" % (name, chi2)
> >>
> >> #--end
> >
> > This should also work.
> >
> >
> >> But this does not seem to work. Can someone help me with the proper
> >> code to extract chisq values from multiple runs?
> >
> > If the script fails with the fix, would you be able to attach a print
> > out of the error message to help in solving where the script failed?
> >
> > Cheers,
> >
> > Edward
>
>
_______________________________________________
relax (http://nmr-relax.com)
This is the relax-users mailing list
[email protected]
To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-users
