Methods and formulas for the response information in Random Forests® Classification

Note

This command is available with the Predictive Analytics Module. Click here for more information about how to activate the module.

The values in the response information table describe the response variable. For a binary response, the table includes the counts of events and non-events. For a multinomial response, the table includes the counts in each class. The table also shows the percentages of each class:

When you add validation with a separate test set to the out-of-bag validation for the analysis, the table displays separate statistics for the test set and the training set.

Class weights

For an analysis with class weights, Minitab provides the option to weight the classes to ensure equal sample sizes across classes. For this type of analysis, the most frequent class has a weight of 1. The following calculation determines the weights for the remaining classes for a binary response, where N represents the sample size:
The following calculation generalizes the class weights for a multinomial response: