TRAININGSAMPLE Subcommand (NAIVEBAYES command)
The TRAININGSAMPLE
subcommand
indicates the method of partitioning the active dataset into training
and test samples. You can specify either a percentage of cases to
assign to the training sample, or you can specify a variable that
indicates whether a case is assigned to the training sample.
- If
TRAININGSAMPLE
is not specified, all cases in the active dataset are treated as training data records. TRAININGSAMPLE
is automatically excluded from factor and covariate variable lists on the NAIVEBAYES command line.
PERCENT Keyword
The PERCENT
keyword specifies
the percentage of cases in the active dataset to randomly assign to
the training sample. All other cases are assigned to the test sample.
The percentage must be a number that is greater than 0 and less than
100. There is no default percentage.
If a weight variable is
defined, the PERCENT
keyword
may not be used.
VARIABLE Keyword
The VARIABLE
keyword specifies
a variable that indicates which cases in the active dataset are assigned
to the training sample. Cases with a value of 1 on the variable are
assigned to the training sample. All other cases are assigned to
the test sample.
- The specified variable may not
be the dependent variable, the weight variable, any variable that
is specified in the factor or covariate lists of the command line,
or any variable that is specified in the factor or covariate lists
of the
FORCE
subcommand. - The variable must be numeric.