| T O P I C R E V I E W |
| kidkolja |
Posted - 01/08/2019 : 6:02:18 PM
Hello there,
I need to fit an error function to some data but can't quite seem to make it. I defined the function using Labtalk (I'm not very experienced in coding) as
y=A*erf(k*(x-x_0))+y_0
Where I specified A, K, x_0 and y_0 as independent parameters
When i made a test run in the function builder it seemed to yield the expected results. But after saving and trying to use it in the simple fit app, it just always returns a straight line across my data, which obviusly isn't the fit I am expecting.
Does somebody know where I'm messing up?
(I'm using OriginPro 2019) |
| 3 L A T E S T R E P L I E S (Newest First) |
| yuki_wu |
Posted - 01/09/2019 : 04:48:54 AM
Hi,
-203 means fit did not converge since mutual dependency exists between parameters. You can try simplifying the function, or fixing several parameter values:
https://www.originlab.com/doc/Origin-Help/The_Reason_Why_Fail_to_Converge#Over-parameterized_functions
Regards,
Yuki
OriginLab
|
| kidkolja |
Posted - 01/09/2019 : 03:12:25 AM
Hey Yuki thank you for helping!
Yes I did set some starting values, but as I fit the function they all take much different end values. My supervisor told me, that the starting values are pretty accurate.
I also forgot to mention, that the fitting yields the error code -203 which seems to say that the function is overparametrized but I don't see how...
|
| yuki_wu |
Posted - 01/09/2019 : 12:29:02 AM
Hi,
Did you try to set the initial parameter values? I mean initial parameter values play a key role in fitting, you should enter “guesstimate” parameter values before starting fitting.
Regards,
Yuki
OriginLab
|
|
|
