The KCLUS Procedure

INPUT Statement

  • INPUT variables <LEVEL= NOMINAL | INTERVAL>;

The INPUT statement specifies the names of the variables to be used in clustering. It names one or more input variables that use common options. If you want to use different options for different variables, you can specify multiple INPUT statements.

To trigger the k-prototypes clustering algorithm for clustering mixed input data that contain both interval and nominal input variables, either you can use two INPUT statements with one INPUT statement that includes the LEVEL=INTERVAL option for interval input variables and another INPUT statement that includes the LEVEL=NOMINAL option for nominal input variables, or you can use a single INPUT statement without the LEVEL= option specified. In the latter case, PROC KCLUS detects the types of the input variables and uses the k-prototypes clustering algorithm if the input variables include both interval and nominal types.

You can include the following option in each INPUT statement:

LEVEL= NOMINAL | INTERVAL

specifies the level of measurement of the variables. For clustering, only interval, binary, and nominal variables are accepted.

By default, LEVEL=INTERVAL for numeric variables, and LEVEL=NOMINAL for categorical character variables.

Last updated: December 21, 2018