ITreeBasedClassifier<T>

Interfaces · AiDotNet.Interfaces

Interface for tree-based classification algorithms.

Implements: IProbabilisticClassifier<T>

Known implementations: DecisionTreeClassifier<T>, ExtraTreesClassifier<T>, GradientBoostingClassifier<T>, RandomForestClassifier<T>

For Beginners

Decision trees work like a flowchart - they ask a series of yes/no questions about features to reach a decision. For example, to classify if an animal is a cat: “Has fur?” (yes) -> “Has whiskers?” (yes) -> “Meows?” (yes) -> “It’s a cat!”

Key properties:

  • MaxDepth: How deep the tree can go (more depth = more complex decisions)
  • Feature importance: Which features were most useful for classification

How It Works

Tree-based classifiers make decisions by learning a series of hierarchical rules from data. They are highly interpretable and can capture non-linear relationships between features.

Properties

Property Summary
Vector<T> FeatureImportances { get; } Gets the feature importance scores computed during training.
int LeafCount { get; } Gets the number of leaf nodes in the tree.
int MaxDepth { get; } Gets the maximum depth of the tree.
int NodeCount { get; } Gets the number of internal (decision) nodes in the tree.