Parameters | Predictive Modeling | Nominalize Continuous Dependent Variables

Nominalize Continuous Dependent Variables
Check this option to use SAS PROC RANK to convert continuous dependent variables into nominal variables of seven categories of equal or nearly equal size and then use the result as a class variable on the right hand side of the model.
This operation increases the number of tests seven-fold but can robustly detect nonlinear associations between continuous dependent variables and predictor variables.
You should specify this option when you suspect the relationship between the response and predictor is nonlinear.
This option is available only when selected statistical testing methods for continuous predictors are specified. Acceptable tests include: T-Test, Asymmetric Loss T-Test, and Quantile Regression T-Test.
Please refer to the SAS PROC RANK documentation for more information.
To Specify This Option:
*
*