Criterion for Stopping Model Selection

Use the drop-down menu to specify how selected models are compared and when the model selection process is terminated.

Available options are described in the table below:

Criteria	Definition 1
None	Selection is not stopped and either all of the predictor or the Maximum Number of Steps are used
ADJRSQ	Adjusted R -square statistic
AIC	Akaike information criterion
AICC	Corrected Akaike information criterion
BIC	Sawa Bayesian information criterion
CP	Mallow C ( p ) statistic
CV	Predicted residual sum of squares with k -fold cross validation
PRESS	Predicted residual sum of squares
SBC	Schwarz Bayesian information criterion
SL	Significance Level If you select SL, you must specify the significance levels for adding and retaining variables .

---

1

Refer to the SAS PROC GLMSELECT documentation for additional details about each of the criteria.

Note : This option is available only when you specify a Model Selection Method other than None .

To Specify the Criterion:

Select the method using the drop-down menu.