| T O P I C R E V I E W |
| erickzhou |
Posted - 09/10/2010 : 04:17:32 AM
Origin Ver 8. and Service Release (Select Help-->About Origin):
Operating System: WinXP
My data is 512*512 size, it is composed by a low frequency background trend and radom noise.
I use "nonlinear surcafe fit" in Origin 8.0 to extract the low frequency background trend. The fitting processe was successful(100), and the "Adj. R-Square" is 0.92173.
Now I have two questions.
First, since the fitting result for background is good, can I say that magnitude of the image grayscale change associated with background trend was substantially greater than the magnitude of the noise?
Second, if the fitting processe was successful(100), is the Adj. R-Square enough and commonly used for statistical description for goodness of fit? |
| 2 L A T E S T R E P L I E S (Newest First) |
| erickzhou |
Posted - 09/13/2010 : 9:27:36 PM
Thank you for your reply.
I have another question. I found in many published papers, "RSS" was often used as the definition of fitting error measures, but to best of my knowledge, I have not found anyone who used "Adj. R-Square" as the definition of fitting error measures. Is the definition of "as the definition of fitting error measures" given out only by Origin software? If it is not, can you give me some reference about those who choose "Adj. R-Square" as the definition of fitting error measures.
Thank you very much! |
| larry_lan |
Posted - 09/13/2010 : 10:26:39 AM
Hi:
For Q1, maybe you can compare two fits?
For Q2, it's really hard to say. Any decision should be base on your research. And there are some values/graphs you can look into.
Thanks
Larry |
|
|
