Author 
Topic 

RiverFalls
USA
26 Posts 
Posted  01/01/2022 : 8:48:15 PM

OriginPro 2022 (64bit) SR1 on Win10 I've just recently tried out the ODR fitting method in OriginPro, and am trying to understand the values of the fitting results.
I made up a simple dataset of x, xerr, y, yerr values and fit it to a straight line equation in the NLSF, as well as using the 'Linear Fit with X error' option. Both routines produced the same fit results (parameters, parameter unc., RSS, etc.). However, when I try to make sense of the RSS value and the residual values, I cannot figure out how those values are being determined. Here is the NLSF fit results:
When I look at the curve tab, I see:
Question 1): How are the "Residual of Sheet1..." values calculated? I cannot find a reference that explains these, and cannot figure them out based on the actual data values (columns C,E) and the values of the fit (columns A and B). For example, the very first row indicates that the 'y' residual should be: 20.04577 = 1.95423, yet column F has something completely different. Question 2): The online information indicates that the RSS value is calculated the same as usual (for LM fits), but I cannot reproduce the value being reported by the ODR fit. I think this must be related to the first question.
Thank you for any help that anyone can provide. Lowell


Castiel
329 Posts 
Posted  01/02/2022 : 10:35:04 AM

quote: Originally posted by RiverFalls
OriginPro 2022 (64bit) SR1 on Win10 I've just recently tried out the ODR fitting method in OriginPro, and am trying to understand the values of the fitting results.
I made up a simple dataset of x, xerr, y, yerr values and fit it to a straight line equation in the NLSF, as well as using the 'Linear Fit with X error' option. Both routines produced the same fit results (parameters, parameter unc., RSS, etc.). However, when I try to make sense of the RSS value and the residual values, I cannot figure out how those values are being determined. Here is the NLSF fit results:
When I look at the curve tab, I see:
Question 1): How are the "Residual of Sheet1..." values calculated? I cannot find a reference that explains these, and cannot figure them out based on the actual data values (columns C,E) and the values of the fit (columns A and B). For example, the very first row indicates that the 'y' residual should be: 20.04577 = 1.95423, yet column F has something completely different. Question 2): The online information indicates that the RSS value is calculated the same as usual (for LM fits), but I cannot reproduce the value being reported by the ODR fit. I think this must be related to the first question.
Thank you for any help that anyone can provide. Lowell
Orthogonal Distance Regression (ODR) is a method that is used to calculate the error perpendicular to the line rather than vertically.
Google odrpack95 for details of the algorithm.

Be The Change
You Want To See
In The World





Topic 


