Forecasting using SAS Forecast Server, SAS/ETS, and more

Proc Model log likelihood

Accepted Solution Solved
Reply
Learner
Posts: 1
Accepted Solution

Proc Model log likelihood

I am estimating a nonlinear regression using proc model and would like to output the log likelihood. Can this be ouputted?


Accepted Solutions
Solution
‎01-31-2017 12:34 PM
SAS Employee
Posts: 25

Re: Proc Model log likelihood

PROC MODEL supports a variety of estimation methods which can be specified using options on the FIT statement.  Most of the estimation methods do not use a log likelihood function in their formulation.  However, for the FIML method, a log likelhood function is evaluated and is reported in the Estimation Summary Statistics table. You can read more about the FIML method used in PROC MODEL here:

http://documentation.sas.com/?docsetId=etsug&docsetVersion=14.2&docsetTarget=etsug_model_sect070.htm...

 

A log likelihood value is also used and reported when you specify your own likelihood using an 'errormodel ~ general' statement, or when you estimate a model using the t distribution.

View solution in original post


All Replies
Solution
‎01-31-2017 12:34 PM
SAS Employee
Posts: 25

Re: Proc Model log likelihood

PROC MODEL supports a variety of estimation methods which can be specified using options on the FIT statement.  Most of the estimation methods do not use a log likelihood function in their formulation.  However, for the FIML method, a log likelhood function is evaluated and is reported in the Estimation Summary Statistics table. You can read more about the FIML method used in PROC MODEL here:

http://documentation.sas.com/?docsetId=etsug&docsetVersion=14.2&docsetTarget=etsug_model_sect070.htm...

 

A log likelihood value is also used and reported when you specify your own likelihood using an 'errormodel ~ general' statement, or when you estimate a model using the t distribution.

☑ This topic is SOLVED.

Need further help from the community? Please ask a new question.

Discussion stats
  • 1 reply
  • 144 views
  • 2 likes
  • 2 in conversation