Efficient Dimensionality Reduction for Canonical Correlation Analysis
Data Structures and Algorithms
2013-05-03 v4 Numerical Analysis
Abstract
We present a fast algorithm for approximate Canonical Correlation Analysis (CCA). Given a pair of tall-and-thin matrices, the proposed algorithm first employs a randomized dimensionality reduction transform to reduce the size of the input matrices, and then applies any CCA algorithm to the new pair of matrices. The algorithm computes an approximate CCA to the original pair of matrices with provable guarantees, while requiring asymptotically less operations than the state-of-the-art exact algorithms.
Cite
@article{arxiv.1209.2185,
title = {Efficient Dimensionality Reduction for Canonical Correlation Analysis},
author = {Haim Avron and Christos Boutsidis and Sivan Toledo and Anastasios Zouzias},
journal= {arXiv preprint arXiv:1209.2185},
year = {2013}
}
Comments
22 pages. 4 figures. To appear in ICML 2013: The 30th International Conference on Machine Learning