IPruneTool - a helper interface class to prune a decision tree.
Any tool which implements the interface should provide two modes for tree pruning:
- automatically find the "best" prune strength by minimizing the error rate on a test sample if SetAutomatic() is called, or if automatic = kTRUE argument is set in CalculatePruningInfo() In this case, the PruningInfo object returned contains the error rate of the optimally pruned tree, the optimal prune strength, and the sequence of nodes to prune to obtain the optimal pruned tree from the original DecisionTree
- a user-provided pruning strength parameter is used to prune the tree, in which case the returned PruningInfo object has QualityIndex = -1, PruneStrength = user prune strength, and PruneSequence is the list of nodes to prune
Definition at line 70 of file IPruneTool.h.