topic Re: LASSO application in ordinal logistic regression in Statistical Procedures

LASSO application in ordinal logistic regression

cksuemoto — Mon, 28 Nov 2016 18:56:47 GMT

Hi all,

I need to implement LASSO for variable selection in the cotext of an ordinal logistic regression.

I found the awesome command PROC GLMSELECT, but it does not apply to my case because my outcome is ordinal.

Please help!

Thanks

Claudia

Re: LASSO application in ordinal logistic regression

Rick_SAS — Mon, 28 Nov 2016 19:54:17 GMT

I wonder if you have looked at the HPGENSELECT procedure, which supports LASSO. You didn't supply any data, but try to modify this code, which is taken from the LOGISTIC documentation for ordinal regression (Caution: untested! Didn't verify that this results in cumulative logit model.)

data Cheese;
   do Additive = 1 to 4;
      do y = 1 to 9;
         input freq @@;
         output;
      end;
   end;
   label y='Taste Rating';
   datalines;
0  0  1  7  8  8 19  8  1
6  9 12 11  7  6  1  0  0
1  1  6  8 23  7  5  1  0
0  0  0  1  3  7 14 16 11
;

title 'Multiple Response Cheese Tasting Experiment';
proc hpgenselect data=Cheese plots=all;
   freq freq;
   class Additive(ref='4');
   model y=Additive / link=logit ;
   selection details=all method=lasso;
run;

Re: LASSO application in ordinal logistic regression

cksuemoto — Tue, 29 Nov 2016 17:16:27 GMT

Dear Rick

Thanks a lot, it seems that it is exactly what I need.

The only problem is that I have SAS 9.3, which does not have this command.

Any other option?

Best,

Claudia