Module tree

Implements a very generic decision tree. This means that the data for an interior node and for leafs can be anything. The splitting criteria of the training set, the features used and the impurity computation can be implemented as required.

Structs

DecisionTree

A decision tree for prediction all kinds of thing a decision tree can predict. A leaf of this tree contains data of type L. This tree uses a TreeFunction of Type F.

TreeParameters

Struct for learning a decision tree

Enums

Binar

There a two cases for a binary tree: Using the left child or the right one. So here we have a Zero or One to distinguish between these cases.

Traits

TreeFunction

A trait for describing the behavior of a learned tree. A Tree which follows this behavior can accept data of type Data and used parameter of type Parameter for interior nodes.

TreeLearnFunctions

A trait for describing the behavior of a tree and the way it should be trained. For training data of type Truth are used to describe the result this tree should have. Data of type LeafParam are saved in leaves. For generating feature where the best one should be chosen for training an interior node an iterator of type ParamIter is used.