Extremely Randomized Trees System

An Extremely Randomized Trees System is an Random Forests System that implements an ExtraTrees Algorithm to solve a ExtraTrees Task.

• AKA: Extra Trees Learning System, Ensemble Extra Trees System, Extra Trees System.
• Context:
  • It ranges from being an Extra Trees Regression System to being a Extra Trees Classification System.
  • …
• Example(s):
  • sklearn.ensemble.ExtraTreesClassifier
  • sklearn.ensemble.ExtraTreesRegressor
• Counter-Example(s):
  • An AdaBoost System.
  • A Bagging System.
  • A Gradient Boosting System.
  • A Random Forests System.
• See: Random Forests Model, Out-of-Bag Error, Partial Permutation, K-Nearest Neighbor Algorithm.

---

---

References

2017

• (Wikipedia, 2017) ⇒ https://en.wikipedia.org/wiki/Random_forest#ExtraTrees Retrieved:2017-11-5.
  • Adding one further step of randomization yields extremely randomized trees, or ExtraTrees. These are trained using bagging and the random subspace method, like in an ordinary random forest, but additionally the top-down splitting in the tree learner is randomized. Instead of computing the locally optimal feature/split combination (based on, e.g., information gain or the Gini impurity), for each feature under consideration, a random value is selected for the split. This value is selected from the feature's empirical range (in the tree's training set, i.e., the bootstrap sample).

---