| Author |
 Topic  |
|
|
Deutscheman
Posts |
 Posted - 10/13/2006 : 10:04:41 AM
|
I fit using non linear curve fit. When I use weighting method one problem appears, because I have not only y-error but also x-errors, so I should take into account both for correct weighting of chi-square. But I did not find it possibility. Real function of weighted chi square in this case
Chi^2=Sum[(y(i)-F(x(i)))/(erry(i)^2-{F(x(i)+errx(i))-F(x(i)-errx(i))/2}^2)]
So, maybee i can change calculation of chi-square on this formula?
But, maybe Origin can do it correctly, then write me, please, how?
Edited by - Deutscheman on 10/13/2006 11:41:06 AM |
|
|
easwar
USA
1965 Posts |
Posted - 10/14/2006 : 1:36:35 PM
|
Hi Deutscheman,
The minimization process, which iteratively finds minimum in chi-square, needs to take the X error into account. This is not supported in the current version.
Easwar
OriginLab
|
 |
|
| |
Topic |
|
|
|
Need help in correct fitting