Class Clustering Variables

Use this field to indicate class variables to be included as predictor variables.

For each level of these variables, a numeric indicator variable of 0s and 1s indicating class membership will be constructed when necessary.

Caution: Specifying a very large number of predictor variables can prohibitively increase execution time. If the number of predictors is greater than, say 5000, it is likely desirable to perform some kind of predictor reduction before running this process. See options on the Predictor Reduction tab. Certain analytical processes, such as Marker-Trait Association and ANOVA are also useful for filtering potential predictor variables and you might consider exploring these as well.

For a large number of predictor variables with a common prefix in their name, use the List-Style Specification of Class Clustering Variables to specify these variables.

To Specify Class Clustering Variables

Specify an Input SAS Data Set.

All of the variables in the input data set are displayed in the Available Variables field.

Examine the list of available variables.

Highlight the desired variables.

Click to add the highlighted variable to the Class Clustering Variable field, as shown below: