Std Error of nonlinear regression coefficients

D

Dolphin72

Guest
#1
Hi There,

I have a commercial NET dll computing nonlinear regression coefficients and their sigma values. However, I would like to compute the confidence interval for each coefficient. I believe that I need to compute first the standard error for each coefficient, but if I use SE=Sigma/SQRT(n) then the obtained values seems to be very different that those calulated by statistical software packages. How do obtained the standard errors correctly? Can someone help?

Thanks for any help in advance

Oliver