Page 1 of 1

Maximum Likelihood Estimate

Posted: Wed Apr 28, 2010 9:18 pm
by joseph
I run LS estimate for Linear Regression but error distibution is not normarlity distribution so i want to use another method to analysis. I think it is Maximum Likelihood Estimate(MLE). How to use MLE for Linear Regession in Eview?

Thanks

Re: Maximum Likelihood Estimate

Posted: Wed Apr 28, 2010 10:28 pm
by EViews Gareth
You need to use the LogL object. It is pretty well documented in the User Guide.

Re: Maximum Likelihood Estimate

Posted: Thu Apr 29, 2010 1:58 am
by joseph
it is complex, so difficult to understand. :(

Re: Maximum Likelihood Estimate

Posted: Thu Apr 29, 2010 5:32 am
by samijo
Isn't normality of the error is a necessary condition for MLE, I might be wrong, but I guess you meant GMM!

Re: Maximum Likelihood Estimate

Posted: Thu Apr 29, 2010 7:02 am
by EViews Gareth
No, normality is not required for MLE.

Re: Maximum Likelihood Estimate

Posted: Thu Apr 29, 2010 10:23 pm
by joseph
Thank you!

Because I want to use various methods to analysis, possible to compare results together.

please give me some idea for estimate methods (linear regression), Thank you!

Joseph