using constraints in nonlinear regression

[Ibrahim Bostan]

Hello,
I am able to estimate the parameters (b1,b2) of the model below. My question is how can i put a constraint so that sum of ( exp({-b1}*lag))*(1-exp(-{b2}*lag))) over lag (1,2,3,4, ....35) is unity.

nl ( dep_var= cited_65_69* i_class_1* i_gyearctng_1*( exp({-b1}*lag))*(1-exp(-{b2}*lag)))

thanks a lot in advance
Ibrahim

[Ibrahim Bostan]

any insights?

[Totonio89]

Hello Ibrahim

Did you find an answer to this question? I know it has been a while but I am also very interested in this question.

Thanks a lot to you or anyone who can help.

Best,

Totonio

[Rich Goldstein]

did you try looking at help for -constraint-? if that doesn't help, tell us specifically what you entered and what Stata responded and why you think it is not what you expected

[Totonio89]

Dear Rich Goldstein.

If I am correct, it is not possible to use the constraints option with nl because you can easily manually constraint the coefficients. However, in this specific case, it is not possible to directly tell stata that the sum should be equal to 1. This is why I am also a bit confused here.

Anyone can help?

Best

Totonio