Programming the statistical procedures from SAS

Generalized Logistic Regression (appeared in Andrew Ng's Machine Language Couresea)

Accepted Solution Solved
Reply
Contributor
Posts: 20
Accepted Solution

Generalized Logistic Regression (appeared in Andrew Ng's Machine Language Couresea)

Dear SAS Gurus,

 

I wantched recently Andrew Ng's Coursea class and found it very fun. I think  many of you found it very interesting.

My questions is Andrew introduced some generalized term of logistic regressions rather than some stepwise selection of variables as follows. 

 

Generalized_term.png

 

Do you know this generalization can be achieved by using any options of proc logistic or any other suitable procedures?

Again, I really thanks to everyone for the help all the time,

Kaz


Accepted Solutions
Solution
‎04-19-2017 10:16 PM
SAS Employee
Posts: 187

Re: Generalized Logistic Regression (appeared in Andrew Ng's Machine Language Couresea)

That looks like it is just an L2 penalty added to the binomial log likelihood resulting in a shrinkage estimation method called ridging. See this note which discusses ways to do shrinkage methods (LASSO, ridging, elastic net) in SAS.

View solution in original post


All Replies
Solution
‎04-19-2017 10:16 PM
SAS Employee
Posts: 187

Re: Generalized Logistic Regression (appeared in Andrew Ng's Machine Language Couresea)

That looks like it is just an L2 penalty added to the binomial log likelihood resulting in a shrinkage estimation method called ridging. See this note which discusses ways to do shrinkage methods (LASSO, ridging, elastic net) in SAS.

Contributor
Posts: 20

Re: Generalized Logistic Regression (appeared in Andrew Ng's Machine Language Couresea)

Dear StatDave

Thanks very much for the advice.
I think the note discuss exactly what I meant to know.

Kaz

##- Please type your reply above this line. Simple formatting, no
attachments. -##
☑ This topic is SOLVED.

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

Discussion stats
  • 2 replies
  • 104 views
  • 3 likes
  • 2 in conversation