10-10-2010 10:05 PM
Hello All,
Thanks for my query answered in the last post. One thing I love about Labview is its ability to allow complex models to be fitted to experimental data with so much ease. i have been using the L-M non-constrained fitting for a long time now. However there is one question which I thought I will like to be answered by Labview professionals themselves.
When the curve fitting is done, does the algorithm always present the minimum value for an inverted parameter. eg. suppose a parameter X has been found to have a value of 0.8 after fitting using L-M algorithm. However in principle several orders of values are possible for X. eg it is not sensitive parameter in the main equation so it can have values of 1000, or 400, or 60, or 1. In this scenario will the algorithm always give an answer as X= 1; that is the minimal solution?
Cheers,
navneet
Solved! Go to Solution.
10-11-2010 09:28 AM
navneet,
The Lev-Mar implementation in LabVIEW does not explicitly try to minimize the magnitude of any parameter. If the parameter has only a minimal impact on the distance between the data and best fit curve, then it will change according to the gradient of the problem. If the parameter does not have any impact on the distance, then you may get an error from the linear solver indicating rank deficiency.
You can detect this behavior by using the Nonlinear Curve Fit Intervals.vi. An interval that is very large compared to the other parameters would indicate a weak dependence on that parameter.
-Jim
05-10-2011 06:51 PM
Thanks Jim.