Dear Sir,
I would like to know how to get the matrix of variance-covariance of parameter estimates using the Maximum likelihood procedure,
I mean after getting the results of the parameters I wan also to know the Variance-Covariance matrix of these parameters, as an extra output of the model estimation.
I have tried to use the following,
y=vcm(_max_FinalHess);
but I don't think its correct!
its so necessary to conduct Andrew's test which needs (f-fe)'V(f-fe), where V: is the varaince-covaraince of the parameter estiamtes
Thank you very much
2 Answers
0
Hi,
If you are using the maximum likelihood package, you should have the following code to call the package in your estimation:
{x,f,g, vc, retcode} = maxlik(data,var,&fct,start);
which basically stores the estimated variance-covariance matrix in the global variable named `vc'.
It will also be printed to the output window by the following code:
maxprt(x,f,g,vc,retcode);
Hope this helps.
Cheers
0
Dear Sir
I would like to thank you very much for answering my question, you saved my day,
Indeed now I can recall the variance-covariance matrix as I want,
yours sincerely
Ghasak
Your Answer
2 Answers
Hi,
If you are using the maximum likelihood package, you should have the following code to call the package in your estimation:
{x,f,g, vc, retcode} = maxlik(data,var,&fct,start);
which basically stores the estimated variance-covariance matrix in the global variable named `vc'.
It will also be printed to the output window by the following code:
maxprt(x,f,g,vc,retcode);
Hope this helps.
Cheers
Dear Sir
I would like to thank you very much for answering my question, you saved my day,
Indeed now I can recall the variance-covariance matrix as I want,
yours sincerely
Ghasak