| T O P I C R E V I E W |
| 73659 |
Posted - 02/24/2006 : 6:13:49 PM
Origin Version (Select Help-->About Origin): 6.0
Operating System: windows XP
I try to fit multiple datasets to the same following nonlinear function
y=f(x, B1, B2),
where x and y are independent and dependent variables, B1 and B2 are parameters with different values for different data sets. I hope that B1 and B2 values for dataset2 are larger than those for dataset1, respectively. However, if I set the following constraints
constant1<B1(dataset1)<B1_2(dataset2)
constant2<B2(dataset1)>B2_2(dataset2),
the constraints of the upper and lower limits, B1_2 and B2_2 in this example, are replaced "permanently" by their initial values, rather than their iterative values as I hope, even though the B1_2 and B2_2 values are updated at every iteration.
How can I fix the problem and obtain the best fitting with the constrains I hope?
|
| 2 L A T E S T R E P L I E S (Newest First) |
| Leo_Li |
Posted - 02/27/2006 : 11:22:15 PM
Hi Wang,
If you are using Options->Constraints dialog page to set your constraints as follows (e.g.):
3.0<B1<B1_2;
1.0<B2<B2_2;
and make sure that Enabled checkbox is checked, then the above constraints will take effect throughout your fitting process.
If the intial values of parameters (before clicking "100 Iter" button) violate the above constraint, an error message box "Error! Constraints are violated during fitting iterations" pops up. Otherwise, the parameters vary under the control of the constraint at every iteration.
Leo
OriginLab Corp. |
| Mike Buess |
Posted - 02/25/2006 : 12:20:29 PM
Like you, I suspect that the NLSF dialog uses the initial constraints throughout the entire fitting session. You can probably get around that with a LabTalk script which changes the constraints after every iteration (or a given number of iterations). The NLSF object has the properties and methods necessary for scripting your fitting session...
Help > LabTalk > Object Reference > Alphabetical Listing... > NLSF
Mike Buess
Origin WebRing Member |
|
|
