| forum.alglib.net http://forum.alglib.net/ |
|
| Linear regression http://forum.alglib.net/viewtopic.php?f=2&t=397 |
Page 1 of 1 |
| Author: | araripe [ Mon Jul 04, 2011 2:21 pm ] |
| Post subject: | Linear regression |
Hi, I am performing some multiple linear regression with Alglib using lrbuild and lrunpack. However, I didn't realize how to obtain the regression coefficient r2 and the standard deviation of each coefficient. Any help? Thank in advance Alexandre Araripe Federal University of Bahia Physics Intitute Brasil |
|
| Author: | Sergey.Bochkanov [ Mon Jul 04, 2011 8:12 pm ] |
| Post subject: | Re: Linear regression |
As for ALGLIB 3.4.0, such functionality is not supported. I understand that it is significant gap in the library, but these functions have lesser priority than other items. I hope that things will get better during second part of the year, when I will get more developers working on ALGLIB. |
|
| Author: | araripe [ Tue Jul 05, 2011 12:36 pm ] |
| Post subject: | Re: Linear regression |
Thank you very much. I will be connected and wainting for such changes. By the way, congratulations for the good job you are done with AlgLib. Alexandre Araripe Federal University of Bahia Physics Intitute Brasil |
|
| Author: | rbeadie [ Thu May 17, 2012 7:27 pm ] |
| Post subject: | Re: Linear regression |
Alexandre, I was trying to solve the same issue, and I believe I have the answer to your problem. The lrreport that is returned includes a covariance matrix. The coefficient standard error is the square root of the diagonal element in that matrix. i.e., the standard error of the jth coefficient is the [j,j] value in the covariance matrix. hope this helps |
|
| Author: | araripe [ Mon Jun 04, 2012 2:24 pm ] |
| Post subject: | Re: Linear regression |
I think this approach will be valuable. Where can I find the math justification for this? Thank you very much. Araripe |
|
| Page 1 of 1 | All times are UTC |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|