Extra Trees

A classifier based on Random Forests, with the difference being that a random feature is selected for the split process. this increases randomness and lowers correlation between features

Notes:

  • Extra Trees use the whole original samples to generate subsets of size N for training ensemble members (Decision trees).