Decision trees
Introduction
Decision trees do recursive partitioning of the data for classification and/or
regression tasks.
DT comes with the risk of overfitting.
- Advantages
- Output is easy to understand
- Can combine numeric and categorical data
- Robust (outliers)
- Fast (after developing the rules)
- Disadvantages
- Prone to overfitting
- Limited to the range of attributes in the training data
- Unstable (small perturbation input -> larger perturbation output)
- Categorical data
Examples
Outgoing relations
- Decision trees is part of Machine learning
- Decision trees is part of Random forest
Incoming relations
- CART is part of Decision trees
- Pruning is part of Decision trees
- Overfitting is part of Decision trees
- Impurity is part of Decision trees
- Node is part of Decision trees
- Root node is part of Decision trees
- Terminal node is part of Decision trees
Contributors
- Sander Mooren