search

doubleplusplus / incremental_decision_tree-CART-Random_Forest

incremental CART decision tree, based on the hoeffding tree i.e. very fast decision tree (VFDT), which is proposed in this paper "Mining High-Speed Data Streams" by Domingos & Hulten (2000). And a newly extended model "Extremely Fast Decision Tree" (EFDT) by Manapragada, Webb & Salehi (2018). Added new implementation of Random Forest
100 stars 28 forks source link

Pruning decision trees #4

Closed charlesxjyang closed 4 years ago

charlesxjyang commented 4 years ago

Would there be any way to implement pre/post pruning of either VFDT/EFDT?

doubleplusplus commented 4 years ago

It's possible to prune VFDT, but I am too lazy to implement the pruning. I suppose you can use an ensemble of VFDT, like a Random Forest style to avoid pruning.

For EFDT, I don't think you need to prune it. Because it's dynamic and can modify itself based on future data instances. Therefore pruning is now necessary.