Author |
Topic  |
|
jmeirlab
Posts |
Posted - 05/15/2006 : 7:21:07 PM
|
I think this is a simple question. I want to get a p-value for an exponential regression I ran, but I can't seem to figure out how to do it. I can get one for a polynomial or linear regression, but not for an exponential decay regression. Can anyone help? Thanks Cass |
|
zachary_origin
China
Posts |
Posted - 05/15/2006 : 11:14:05 PM
|
quote:
I think this is a simple question. I want to get a p-value for an exponential regression I ran, but I can't seem to figure out how to do it. I can get one for a polynomial or linear regression, but not for an exponential decay regression. Can anyone help? Thanks Cass
Hi, There is a sample : *origin\samples\analysis\curve fitting\Exp Decay.opj (*origin is the Origin installation) You can take a look at it.
Hope it helps,
Zachary OriginLab GZoffice. |
 |
|
jmeirlab
Posts |
Posted - 05/16/2006 : 12:38:06 PM
|
Hi Thanks for the response. However, I looked at the sample, but as far as I can tell, it just shows how to plot exponential decay. The p-value in the window is from a linear regression of a different data file. Any other suggestions |
 |
|
jmeirlab
Posts |
Posted - 05/16/2006 : 12:38:09 PM
|
Hi Thanks for the response. However, I looked at the sample, but as far as I can tell, it just shows how to plot exponential decay. The p-value in the window is from a linear regression of a different data file. Any other suggestions? |
 |
|
azhu
China
4 Posts |
Posted - 05/17/2006 : 02:24:19 AM
|
Hi,
I assume what you need is the P-value in ANOVA table. Well, in NLSF, it can be computed as following,
F statstic = (MSS of Model) / (Chi^2/dof) where MSS of Model is computed as: (Sum((y-y_ave)^2) - Chi^2)/(k-1). k is number of parameters and y_ave is mean of y
P-value = 1 - invf(F statistic, k-1, n-k)
Thanks Echo Originlab GZOffice
Edited by - azhu on 05/17/2006 02:25:31 AM
Edited by - azhu on 05/17/2006 02:44:25 AM |
 |
|
zachary_origin
China
Posts |
Posted - 05/17/2006 : 08:40:56 AM
|
 |
 |
|
|
Topic  |
|