Hi Fernando,
The `fgls` function does not directly provide the option to return the innovations covariance matrix used in estimation. However, you can compute the innovations covariance matrix manually using the residuals obtained from the `fgls` estimation.
Here's an example of how you can compute the innovations covariance matrix:
Mdl1 = fgls(Mdl, y);
residuals = Mdl1.Residuals.Raw;
innovationsCov = cov(residuals);
Thanks,
Balaji
