Page 1 of 1

NLS estimation problem

Posted: Thu Jun 27, 2013 4:55 am
by aydam6
Hi,

I want to estimate the equation with NLS method:

C_FUEL=C(1)*TC+((1-C(1))*28.6*C(2)-C(1)*38.9*C(3)-C(1)*40.9*C(4))*N

The estimation out put gives me coef. but it does not have prob or t-statistic and it gives me the 'WARNING: Singular covariance - coefficients are not unique'

C(2) and C(3) and C(4) must be positive parameters but i can't add this restriction to equation.

Thank you

aida

Re: NLS estimation problem

Posted: Thu Jun 27, 2013 6:24 am
by startz
I believe c(2) and c(4) are perfectly multicollinear.