Saved in:
Bibliographic Details
Main Authors: Li, Hongyi, Lin, Han, Xu, Jun
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.05371
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917451331010560
author Li, Hongyi
Lin, Han
Xu, Jun
author_facet Li, Hongyi
Lin, Han
Xu, Jun
contents Oblique decision trees combine the transparency of trees with the power of multivariate decision boundaries, but learning high-quality oblique splits is NP-hard, and practical methods still rely on slow search or theory-free heuristics. We present the Hinge Regression Tree (HRT), which reframes each split as a non-linear least-squares problem over two linear predictors whose max/min envelope induces ReLU-like expressive power. The resulting alternating fitting procedure is exactly equivalent to a damped Newton (Gauss-Newton) method within fixed partitions. We analyze this node-level optimization and, for a backtracking line-search variant, prove that the local objective decreases monotonically and converges; in practice, both fixed and adaptive damping yield fast, stable convergence and can be combined with optional ridge regularization. We further prove that HRT's model class is a universal approximator with an explicit $O(δ^2)$ approximation rate, and show on synthetic and real-world benchmarks that it matches or outperforms single-tree baselines with more compact structures.
format Preprint
id arxiv_https___arxiv_org_abs_2602_05371
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Hinge Regression Tree: A Newton Method for Oblique Regression Tree Splitting
Li, Hongyi
Lin, Han
Xu, Jun
Machine Learning
Oblique decision trees combine the transparency of trees with the power of multivariate decision boundaries, but learning high-quality oblique splits is NP-hard, and practical methods still rely on slow search or theory-free heuristics. We present the Hinge Regression Tree (HRT), which reframes each split as a non-linear least-squares problem over two linear predictors whose max/min envelope induces ReLU-like expressive power. The resulting alternating fitting procedure is exactly equivalent to a damped Newton (Gauss-Newton) method within fixed partitions. We analyze this node-level optimization and, for a backtracking line-search variant, prove that the local objective decreases monotonically and converges; in practice, both fixed and adaptive damping yield fast, stable convergence and can be combined with optional ridge regularization. We further prove that HRT's model class is a universal approximator with an explicit $O(δ^2)$ approximation rate, and show on synthetic and real-world benchmarks that it matches or outperforms single-tree baselines with more compact structures.
title Hinge Regression Tree: A Newton Method for Oblique Regression Tree Splitting
topic Machine Learning
url https://arxiv.org/abs/2602.05371