Extremely Fast Decision Tree
This work provides a higher accuracy replacement for Hoeffding Tree in scenarios like streaming data, benefiting researchers and practitioners in machine learning, though it is incremental in nature.
The authors tackled the problem of improving incremental decision tree learning by introducing Hoeffding Anytime Tree, which achieves significantly superior prequential accuracy on most large UCI classification datasets compared to the state-of-the-art Hoeffding Tree, with only a small additional computational cost.
We introduce a novel incremental decision tree learning algorithm, Hoeffding Anytime Tree, that is statistically more efficient than the current state-of-the-art, Hoeffding Tree. We demonstrate that an implementation of Hoeffding Anytime Tree---"Extremely Fast Decision Tree", a minor modification to the MOA implementation of Hoeffding Tree---obtains significantly superior prequential accuracy on most of the largest classification datasets from the UCI repository. Hoeffding Anytime Tree produces the asymptotic batch tree in the limit, is naturally resilient to concept drift, and can be used as a higher accuracy replacement for Hoeffding Tree in most scenarios, at a small additional computational cost.