Boosting Revisited: Benchmarking and Advancing LP-Based Ensemble Methods
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915685220745216 |
|---|---|
| author | Akkerman, Fabian Ferry, Julien Artigues, Christian Hebrard, Emmanuel Vidal, Thibaut |
| author_facet | Akkerman, Fabian Ferry, Julien Artigues, Christian Hebrard, Emmanuel Vidal, Thibaut |
| contents | Despite their theoretical appeal, totally corrective boosting methods based on linear programming have received limited empirical attention. In this paper, we conduct the first large-scale experimental study of six LP-based boosting formulations, including two novel methods, NM-Boost and QRLP-Boost, across 20 diverse datasets. We evaluate the use of both heuristic and optimal base learners within these formulations, and analyze not only accuracy, but also ensemble sparsity, margin distribution, anytime performance, and hyperparameter sensitivity. We show that totally corrective methods can outperform or match state-of-the-art heuristics like XGBoost and LightGBM when using shallow trees, while producing significantly sparser ensembles. We further show that these methods can thin pre-trained ensembles without sacrificing performance, and we highlight both the strengths and limitations of using optimal decision trees in this context. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_18242 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Boosting Revisited: Benchmarking and Advancing LP-Based Ensemble Methods Akkerman, Fabian Ferry, Julien Artigues, Christian Hebrard, Emmanuel Vidal, Thibaut Machine Learning Despite their theoretical appeal, totally corrective boosting methods based on linear programming have received limited empirical attention. In this paper, we conduct the first large-scale experimental study of six LP-based boosting formulations, including two novel methods, NM-Boost and QRLP-Boost, across 20 diverse datasets. We evaluate the use of both heuristic and optimal base learners within these formulations, and analyze not only accuracy, but also ensemble sparsity, margin distribution, anytime performance, and hyperparameter sensitivity. We show that totally corrective methods can outperform or match state-of-the-art heuristics like XGBoost and LightGBM when using shallow trees, while producing significantly sparser ensembles. We further show that these methods can thin pre-trained ensembles without sacrificing performance, and we highlight both the strengths and limitations of using optimal decision trees in this context. |
| title | Boosting Revisited: Benchmarking and Advancing LP-Based Ensemble Methods |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2507.18242 |