what is the difference between PROC LOGISTIC and PROC GLMSELECT?

[fannavivian]

I am new to SAS/STAT, and I am wondering what is the difference between PROC LOGISTIC and PROC GLMSELECT? The SAS syntax are very similar: both of them can run logistic regression models, both of them can have specific selection method (FORWARD, BACKWARD, STEPWISE), and both of them can be used to score a new dataset. The only difference I observe now is that PROC GLMSELECT can only be used to score a continuous response variable, while PROC LOGISTIC can be used to score binary response variables (eg: specify the event=1). Am I right on this? Is there any other difference that I should keep an eye on? 

[Reeza]

GLMSELECT works on a larger range of models than Logistic Regression. A model from GLMSELECT does not have to be a LOGISTIC regression. 

I think your last statement is backwards? Logistic works on binary or categorical variables, while GLM can be continuous. The models can be very different. I also believe LOGISTIC will have some more specific features such as AUC & ROC curves but I haven't checked that out. The usual comparison is between GLM or REG and GLMSELECT. 

[fannavivian]

Thanks for your reply! Can you give me an example when PROC GLMSELECT should be used rather than PROC LOGISTIC, or vice versa?

[Reeza]

@PGStats Is correct, GLMSELECT doesn't do Logistic Regression. 

Read the Overview for a comparison of the PROCS. 

http://support.sas.com/documentation/cdl/en/statug/68162/HTML/default/viewer.htm#statug_glmselect_ov...

The GLMSELECT procedure performs effect selection in the framework of general linear models. A variety of model selection methods are available, including the LASSO method of Tibshirani (1996) and the related LAR method of Efron et al. (2004). The procedure offers extensive capabilities for customizing the selection with a wide variety of selection and stopping criteria, from traditional and computationally efficient significance-level-based criteria to more computationally intensive validation-based criteria. The procedure also provides graphical summaries of the selection search.

The GLMSELECT procedure compares most closely to REG and GLM. The REG procedure supports a variety of model-selection methods but does not support a CLASS statement. The GLM procedure supports a CLASS statement but does not include effect selection methods. The GLMSELECT procedure fills this gap. GLMSELECT focuses on the standard independently and identically distributed general linear model for univariate responses and offers great flexibility for and insight into the model selection algorithm. GLMSELECT provides results (displayed tables, output data sets, and macro variables) that make it easy to take the selected model and explore it in more detail in a subsequent procedure such as REG or GLM.

[Reeza]

I think PROC GENMOD and possibly GLM can also be used to run a logistic regression model. 

You'll quickly find there are many ways to do one thing in SAS. 

[PGStats · Accepted solution]

As far as I know, glmselect does not do logistic regression. GLM stands for general linear models which do not include the logistic.

Both procedures offer variable selection, but only glmselect offers advanced effect selection and the LAR and LASSO methods. 

[Ksharp]

PROC HPGENSELECT can do logistic .