Model Selection via the VC-Dimension

TL;DR

This paper introduces a new objective function to estimate the VC-dimension for model selection in regression, demonstrating its consistency and superior performance over existing methods across diverse datasets.

Contribution

It proposes a novel estimator for VC-dimension tailored for regression model selection, validated through consistency proofs and empirical comparisons.

Findings

Estimator is consistent across datasets

Performs better than seven existing techniques

Effective for various data types

Abstract

We derive an objective function that can be optimized to give an estimator of the Vapnik- Chervonenkis dimension for model selection in regression problems. We verify our estimator is consistent. Then, we verify it performs well compared to seven other model selection techniques. We do this for a variety of types of data sets.