VARIABLES Subcommand (OPTIMAL BINNING command)
The VARIABLES
subcommand specifies
the guide variable (if applicable) and one or more binning input variables.
It can also be used to save new variables containing the binned values.
GUIDE=variable. Guide variable. The bins formed by supervised
binning methods are "optimal" with respect to the specified guide
variable. You must specify a guide variable to perform MDLP (CRITERIA METHOD = MDLP
) or the hybrid method
(CRITERIA PREPROCESS = EQUALFREQ METHOD =
MDLP
). This option is silently ignored if it is specified
when the equal frequency method (CRITERIA
METHOD = EQUALFREQ
) is in effect. The guide variable may
be numeric or string.
BIN=varlist. Binning input variable list. These are the variables to be binned. The variable list must include at least one variable. Binning input variables must be numeric.
SAVE = NO | YES (INTO = new varlist). Create new variables containing binned
values. By default, the procedure does not create any
new variables (NO
). If YES
is specified, variables containing the
binned values are saved to the active dataset. Optionally, specify
the names of the new variables using the INTO
keyword.
- The number of variables specified on the
INTO
list must equal the number of variables on theBIN
list. All specified names must be valid variable names. Violation of either of these rules results in an error. - If
INTO
is omitted, new variable names are created by concatenating the guide variable name (if applicable) and an underscore '_', followed by the binning input variable name and an underscore, followed by 'bin'. For example,/VARIABLES GUIDE=E BIN=F G SAVE=YES
will generate two new variables: E_F_bin and E_G_bin.