Store prediction statistics for CART® Classification

Run Stat > Predictive Analytics > CART® Classification. Click Predict > Storage.

You can save statistics from your analysis to the worksheet so that you can use them in other analyses, graphs, and macros. Minitab stores the selected statistics after the last column of data. The names of the storage columns end with a number that increases as you store the same statistic multiple times.

Terminal node ID
Minitab stores the terminal node identification number in the column TNodeId[0] for a tree that you create from the menu. For an alternative tree, the value represents the difference from the initial tree in the results. For example, if you select a tree that is 4 smaller than the initial tree in the results, then the column is TNodeId[–4].
Predicted response level
Minitab stores the predicted class in the column TClass[0] for a tree that you create from the menu. For an alternative tree, the value represents the difference from the initial tree in the results. For example, if you select a tree that is 4 smaller than the initial tree in the results, then the column is TClass[–4].
Predicted probability of each response level
For a binary response, Minitab stores the predicted event probability in the column PEvent and the predicted nonevent probability in the column PNonEvent.
  • PEvent[0] contains the predicted event probability for a tree that you create from the menu.
  • PNonEvent[0] contains the predicted nonevent probability for a tree that you create from the menu.
For a multinomial response, Minitab stores the predicted probability for each level in the column PClass.
  • PClass[0](1) represents a tree that you create from the menu, response = 1.
  • PClass[0](2) represents a tree that you create from the menu, response = 2.
  • PClass[0](3) represents a tree that you create from the menu, response = 3.
  • PClass[0](4) represents a tree that you create from the menu, response = 4.
For an alternative tree, the value represents the difference from the initial tree in the results. For example, if you select a tree that is 4 smaller than the initial tree, then the value for the fourth class of a multinomial tree is PClass[–4](4).