Nystrom Approximation for Sparse Kernel Methods: Theoretical Analysis and Empirical Evaluation

Authors

  • Zenglin Xu University of Electronic Science and Technology of China
  • Rong Jin Michigan State University
  • Bin Shen Purdue University
  • Shenghuo Zhu Alibaba Group

DOI:

https://doi.org/10.1609/aaai.v29i1.9626

Keywords:

nystrom approximation, kernel methods, regression

Abstract

Nystrom approximation is an effective approach to accelerate the computation of kernel matrices in many kernel methods. In this paper, we consider the Nystrom approximation for sparse kernel methods. Instead of relying on the low-rank assumption of the original kernels, which sometimes does not hold in some applications, we take advantage of the restricted eigenvalue condition, which has been proved to be robust for sparse kernel methods. Based on the restricted eigenvalue condition, we have provided not only the approximation bound for the original kernel matrix but also the recovery bound for the sparse solutions of sparse kernel regression. In addition to the theoretical analysis, we also demonstrate the good performance of the Nystrom approximation for sparse kernel regression on real world data sets.

Downloads

Published

2015-02-21

How to Cite

Xu, Z., Jin, R., Shen, B., & Zhu, S. (2015). Nystrom Approximation for Sparse Kernel Methods: Theoretical Analysis and Empirical Evaluation. Proceedings of the AAAI Conference on Artificial Intelligence, 29(1). https://doi.org/10.1609/aaai.v29i1.9626

Issue

Section

Main Track: Novel Machine Learning Algorithms