| Author |
 Topic  |
|
|
keex
Posts |
 Posted - 04/21/2012 : 06:25:34 AM
|
Origin Ver. and Service Release (Select Help-->About Origin): 8.1 SR3
Operating System: W7
Hello everyone,
I have a little problem here with data fitting which I can't seem to overcome. I'll first try to explain the situation:
I am trying to fit six curves at once with a single function in the concetanate mode (one parameter set for all curves). So far no problem. The thing is, I need a parameter in the fit function that takes a specific known value for each of the six curves. For example for curve #1 it should have the value 0.85, for curve #2 it should have 4.4 and so on. My fit function should look something like
y = A*x^B + C*x*"Fixedvaluepercurve"
with A,B,C=free parameters, and "Fixedvaluepercurve" being the variable parameter depending on the dataset.
I hope i could make it clear what my problem is. Do you know a way to kind of make a table or define attributes of curves that can then be included in the fit?
Thank you very much in advance,
Best regards,
Keex |
|
|
easwar
USA
1964 Posts |
Posted - 04/22/2012 : 3:12:34 PM
|
Hi Keex,
When fitting in concatenate mode, Origin simply combines all datasets internally into a single dataset, and then that is fit with the chosen function.
With your requirements, perhaps the following may work:
1> Append all your raw datasets into one X,Y dataset (two columns) manually
2> Create a third dataset/column by scaling your individual datsets with the desired scaling factor and then appending them
3> Define a fitting function with two independent variables, where the first indep var is assigned the raw appended values, and then 2nd indep var is assigned the scaled appended value.
Easwar
OriginLab |
 |
|
|
keex
Posts |
Posted - 04/23/2012 : 03:55:48 AM
|
Thank you easwar for your answer. I think I found, that the "Global Fit" method is more suitable to my problem and I can just input the fixed values there for each dataset.
That said I have another followup question:
Is there a way to set parameters to "Shared" in the parameter initialization or somewhere else? I have to try out serveral different fit functions and after every change I have to manually set every parameter to "Shared" again, which gets a bit tedious ;-)
Thanks again.
Greetings,
Keex |
|
|
|
easwar
USA
1964 Posts |
Posted - 04/23/2012 : 09:54:39 AM
|
Hi Keex,
Did you set the fitting mode to be "Global Fit" by changing the "Multi-Data Fit Mode" drop-down under "Data Selection"? Once you do that, the "Parameters" tab should show a new column called "Share" where you can check the check boxes to share the same parameter across multiple datasets.
Note that you can also add constraints on parameters. Click on the "Code" tab and then select "Constraints" from the list, and then look at the "Hints" tab in lower section of the dialog for further instructions on notation.
Once you have set sharing or constraints, you can re-save your FDF and the information will be saved so next time you re-load the function, the settings will be there.
Easwar
OriginLab
|
|
|
| |
Topic |
|
|
|
Concatenate Fitting with fixed parameter per curve