In the Discriminant platform, specifying a validation set to use depends on the version of JMP. In standard JMP, you can specify a validation set by excluding the rows that form the validation set. Select the rows that you want to use as your validation set and then select Rows > Exclude/Unexclude. The unexcluded rows are treated as the training set.
In JMP Pro, you can specify a Validation column in the Discriminant launch window. A validation column must have a numeric data type and should contain at least two distinct values.
Note the following:
• If the column contains two values, the smaller value defines the training set and the larger value defines the validation set.
• If the column contains three values, the values define the training, validation, and test sets in order of increasing size.
• If the column contains four or more distinct values, only the smallest three values and their associated observations are used to define the training, validation, and test sets, in that order.
When a validation set is specified, the Discriminant platform performs the following steps:
• Models are fit using the training data.
• The Stepwise Variable Selection option gives the Validation Entropy RSquare and Validation Misclassification Rate statistics for the model. See Statistics and Entropy RSquare for Validation and Test Sets.
• The Discriminant Scores report shows an indicator identifying rows in the validation and test sets.
• The Score Summaries report shows actual by predicted classifications for the training, validation, and test sets.