| Author |
 Topic  |
|
|
coolingc
1 Posts |
 Posted - 09/04/2012 : 6:57:37 PM
|
Hey all-
I am working on fitting a nonlinear curve function in Origin to a set of data I've taken. The form of the equation is
y = M* (1- (A/x^(1/2))) + K
Where x and y are my data variables, and I am fitting for M, A and K. For some reason, Origin doesn't fit when the exponent attached to the x is less than 1 (ie a fraction).
Often, I get the error ' Fit did not converge - mutual dependency exists between parameters.
You may have overparameterized the fitting function. Fixing one of them may eliminate this problem.'
What can I do to alleviate this issue? Infinite thanks!
|
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 09/05/2012 : 09:45:38 AM
|
Hi coolingc,
I'm not sure, but could you first please check IF your fit function type is OriginC or else. If it is of OriginC, the "1/2" term in your formula becomes 0 because of the integer calculation. Then, eventually your looses the independent variable, and will never converge. If this is the case, you can write 0.5 (as real) or (1.0/2.0) instead of 1/2.
--Hideo Fujii
OriginLab
|
Edited by - Hideo Fujii on 09/05/2012 10:43:34 AM |
 |
|
| |
Topic |
|
|
|
Overparameterized Nonlinear Curve Fit