| Author |
 Topic  |
|
|
Bobby64
Spain
5 Posts |
 Posted - 02/08/2012 : 2:26:49 PM
|
Origin Ver. and Service Release (Select Help-->About Origin): 6 +8.5
Operating System: Windows
Hi,
I have a question related to the non linear fitting procedure with Origin. I guess it must be basic for all mathematicians, but I have to admit that my knowledge in this field is a bit limited...
Here is my question:
I run some non linear fittings of experimental data points using a user defined function. When the fitting iterations are done, and the X-square is reduced, I have a final result showing the values of all the fitted parameters with the associated standard errors.
So, my question is to know whether this standard error calculated by Origin corresponds to 1-sigma or 2-sigma of standard deviation ?
I had a look to the help, but, to be honnest, I didn't find anything so specific... For those who have the answer, could you give where it is written is the help manual?
Many thanks in advance for your help,
Bobby64 |
|
|
Hideo Fujii
USA
1582 Posts |
Posted - 02/08/2012 : 5:00:15 PM
|
Hi Bobby64,
I'm not a mathematician nor a statistician, but my basic understanding is that Standard Error(SE) (of Meam, SEM) is, for calculation of the confident interval, a reduced, or normalized value of Standard Deviation(SD) - SD divided by a square root of sample size. In your term, if I dare to say, SE is 1-sigman reduced SD, thus +/-1.96*SE is the 95% range (1.96 for large N, otherwise t95 value).
Origin's NLFit has also an option to output the confident intervals, LCL and UCL of a given confidence level.
You can find the details of SE of estimated parameters in the Origin help at:
UserGuide> Regression and Curve Fitting> Nonlinear Curve Fitting> The Fit Results> Parameter Standard Errors
(Please don't ask me the meaning of the formula, or go... :"-) )
Hope the above is a good start for you.
--Hideo Fujii
OriginLab
|
 |
|
|
Bobby64
Spain
5 Posts |
Posted - 02/10/2012 : 03:32:45 AM
|
Hi Hideo Fujii,
many thanks for your answer, it helps me a lot.
Cheers
Bobby64 |
|
|
| |
Topic |
|
|
|
Non linear curve fitting -