Random Forest algorithm consistency through grafting onto CART.
Abstract
Introduction to Random Forests and their applications.
Theoretical exploration of consistency in Random Forests.
Comparison of different variants addressing method shortcomings.
Empirical application on Boston Housing dataset.
Experiments on simulated data for performance evaluation.
Role of feature selection in Grafted Trees and Centered Forests.
Theoretical results on the role of αn in Grafted Trees and Generalized Grafting.
Sparse setting analysis under Assumption 2.
Conclusion on the suitability of Grafted Trees for prediction settings.
Customize Summary
Rewrite with AI
Generate Citations
Translate Source
To Another Language
Generate MindMap
from source content
Visit Source
arxiv.org
Grafting
Stats
Random Forests are consistent with L2 consistency guarantee regardless of distribution.
Grafted Trees outperform Centered Forests with a test error of 11.23 compared to 26.45.