Revisiting Chebyshev Polynomial and Anisotropic RBF Models for Tabular Regression

Researchers developed smooth-basis regression models including anisotropic RBF networks and Chebyshev polynomial regressors that compete with tree ensembles in tabular regression tasks. Testing across 55 datasets showed these models achieve similar accuracy to tree ensembles while offering better generalization properties and gradual prediction surfaces suitable for optimization applications.

• →Pre-trained transformers ranked first in accuracy but face deployment constraints due to GPU requirements and latency issues in CPU-based industrial settings.
• →Smooth-basis models and tree ensembles showed statistically tied performance on accuracy across 55 regression datasets.
• →Smooth models demonstrated tighter generalization gaps compared to tree ensembles, indicating better robustness.
• →Three new scikit-learn-compatible packages were released for anisotropic RBF networks, Chebyshev polynomial regressors, and smooth-tree hybrids.
• →Researchers recommend including smooth-basis models in candidate pools, especially when applications benefit from gradual prediction variations.