Ensembles of Kernel Predictors

TL;DR

This paper investigates ensemble methods for combining multiple kernel predictors, providing new theoretical guarantees, a novel learning algorithm, and empirical validation across various datasets.

Contribution

It introduces a comprehensive analysis of ensemble kernel predictors, including theoretical bounds, a new learning algorithm, and empirical comparisons with existing kernel learning methods.

Findings

Theoretical guarantees based on Rademacher complexity.

Effective ensemble learning algorithm demonstrated.

Empirical results show competitive performance.

Abstract

This paper examines the problem of learning with a finite and possibly large set of p base kernels. It presents a theoretical and empirical analysis of an approach addressing this problem based on ensembles of kernel predictors. This includes novel theoretical guarantees based on the Rademacher complexity of the corresponding hypothesis sets, the introduction and analysis of a learning algorithm based on these hypothesis sets, and a series of experiments using ensembles of kernel predictors with several data sets. Both convex combinations of kernel-based hypotheses and more general Lq-regularized nonnegative combinations are analyzed. These theoretical, algorithmic, and empirical results are compared with those achieved by using learning kernel techniques, which can be viewed as another approach for solving the same problem.