| T O P I C R E V I E W |
| polop44 |
Posted - 07/16/2002 : 06:08:18 AM
Hi!
I fitted data points with the non linear fit curve function (with an personal defined function). I got a set of parameter values.
When I checked these values by plotting the function values (not using the non linear curve fit function but the equation values set in a column), the curve does not correspond (and by far!).
Do you have any idea about that?
Thanks for your reply!
|
| 2 L A T E S T R E P L I E S (Newest First) |
| Hideo Fujii |
Posted - 01/04/2006 : 10:44:54 AM
Please note that the above link in the Easwar's message has been changed to the following:
http://www.originlab.com/www/support/resultstech.aspx?ID=43&language=English
--Hideo
|
| easwar |
Posted - 07/16/2002 : 08:52:15 AM
Hello,
The parameters from the fit are stored in the NLSF object as nlsf.p1, nlsf.p2 etc. When you try to reproduce your fit results independently using an equation, you should use these NLSF object properties instead of the values reported in fit results pasted on the graph, for instance.
For example, if your function is:
y=a+b*x+c*x^2
then use the equation:
nlsf.p1+nlsf.p2*col(a)+nlsf.p3*col(a)^2
to compute the values, rather than plugging in the reported values. The two can differ in extremes cases such as polynomial functions of higher orders.
The parameter values from the fit are reported with a preset precision (significant figures). This setting can be changed. See the following knowledge base item:
http://www.originlab.com/www/support/resultstech.asp?ItemID=43&VersionID=All&TitleID=Adding+significant+figures+to+fitting+results+in+the+NLSF&CatID=Fitting&TypeID=Tip
Easwar
OriginLab.
Edited by - easwar on 07/16/2002 08:53:32 |
|
|
