| Author |
 Topic  |
|
|
Tashira
3 Posts |
 Posted - 11/26/2015 : 05:09:33 AM
|
Dear all,
I fitted my measurement data (two diagramms with three independent datasets each) with Modell ExpGro1.
For one diagramm I got for each dataset a correlation coefficient of 1.
For the second diagramm I got for each dataset a correlation coefficient and a standard error of 0. Here I masked one datapoint in each dataset.
Both resulsts seem not plausible.
Does anyone know, what to do?
Thanks in advance
best regards,
Tashira
|
|
|
Echo_Chu
China
Posts |
Posted - 11/27/2015 : 04:15:19 AM
|
Hi, Tashira
Hard to say anything now base on your description.
Could you show us your data, your fitting result and correlation coefficient results? It should also help to show us the graph with data and fitted curves.
Echo
OriginLab Corp. |
 |
|
|
Tashira
3 Posts |
Posted - 12/01/2015 : 06:40:45 AM
|
Hi Echo,
thanks for your answer. Here I send you my data:
NaCl:
c(NaCl) in mM D R µ
0 5,90E-02 9,51E-01 3,01E-04
5 5,58E-02 9,95E-01 1,08E-03
10 4,03E+00 6,91E+00 1,00E-10
20 1,33E+01 9,37E+01 4,30E-01
CaCl2:
c(CaCl) in mM D R µ
0 5,90E-02 9,51E-01 3,01E-04
0,2 6,96E-02 1,02E+00 1,08E-03
0,4 6,64E-01 2,34E+00 1,67E-03
0,6 2,05E+01 1,83E+02 6,70E-01
and the fitting results:
I hope, you can help me.
Best regards
Tashira |
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 12/01/2015 : 4:34:43 PM
|
Hi Tashira,
It seems the fitting failed as the ExpGro1 model couldn't
reach converged within the given tolerance to your data.
Can you try to change the tolerance in NLFit tool at:
Settings tab> Advanced(in the left panel)> Fit Control> Iterations> Tolerance(in the right panel) to e.g., 1E-4 ?
--Hideo Fujii
OriginLab |
|
|
|
Tashira
3 Posts |
Posted - 12/18/2015 : 05:34:02 AM
|
Hi Hideo Fujii,
thanks for your tip.
I set the tolerance to 0.01 and the max. number of iterations to 100.
Therby, I got for D(CaCl2) and R(CaCl2)a correllation coefficient of 0.99999 and 0.99979, resp.
Unfortunately, this change of fit control does not affect all NaCl-Curves and µ(CaCl2). A further reduction of tolerance and max. number of iterations does not help, too.
I wonder, why visual the fitting curves match relatively good to the exp. data, but the correlation coefficients are so strange.
Do you have another idea, what I can do?
Thanks in advance.
Best regards
Tashira
|
|
|
| |
Topic |
|
|
|
correlation coefficinet