[Gretl-users] problem for retrieving internal variables using NLS ?

[Allin Cottrell]

On Fri, 8 Sep 2006, Arnaud Bervas wrote:

> Isn't there a problem for saving the coefficients of NLS 
> estimations in a loop? No problem at all for OLS, but couldn't 
> do the same with NLS

With NLS, you've already got all the coefficients saved!

The parameters have to be created as scalar variables before 
specifying the NLS equation, and after estimation those 
variables contain the final estimates.  If you want to save
successive estimates, you could store them into a matrix.

Allin.