Smooth-basis models challenge tree ensembles in tabular regression

Researchers have revisited Chebyshev polynomial and Anisotropic RBF models for tabular regression, developing new implementations and comparing them against tree ensembles and transformers. While transformers showed higher accuracy, their computational demands limit their applicability. Among CPU-viable models, the smooth-basis models performed comparably to tree ensembles in accuracy but demonstrated better generalization. The study recommends including smooth-basis models in the selection process for tabular regression tasks, especially when gradual prediction variation and tighter generalization are beneficial. AI

IMPACT Smooth-basis models offer a competitive alternative to tree ensembles for tabular data, particularly when generalization is key.

RANK_REASON The cluster contains an academic paper detailing new models and benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• Anisotropic RBF
• arXiv
• CatalyzeX
• Chebyshev polynomial
• DagsHub
• Hugging Face
• Luciano Gerber
• scikit-learn
• tabular regression
• transformer

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →