English

A kernel method for canonical correlation analysis

Machine Learning 2007-05-23 v2 Computer Vision and Pattern Recognition

Abstract

Canonical correlation analysis is a technique to extract common features from a pair of multivariate data. In complex situations, however, it does not extract useful features because of its linearity. On the other hand, kernel method used in support vector machine is an efficient approach to improve such a linear method. In this paper, we investigate the effectiveness of applying kernel method to canonical correlation analysis.

Keywords

Cite

@article{arxiv.cs/0609071,
  title  = {A kernel method for canonical correlation analysis},
  author = {Shotaro Akaho},
  journal= {arXiv preprint arXiv:cs/0609071},
  year   = {2007}
}

Comments

Full version of paper presented in IMPS2001 (International Meeting of Psychometric Society) 2007-Feb-14: typos in equations (23) and (24) in page 3 of the first version have been corrected