English

Multiplicative Algorithm for Orthgonal Groups and Independent Component Analysis

Machine Learning 2007-05-23 v1

Abstract

The multiplicative Newton-like method developed by the author et al. is extended to the situation where the dynamics is restricted to the orthogonal group. A general framework is constructed without specifying the cost function. Though the restriction to the orthogonal groups makes the problem somewhat complicated, an explicit expression for the amount of individual jumps is obtained. This algorithm is exactly second-order-convergent. The global instability inherent in the Newton method is remedied by a Levenberg-Marquardt-type variation. The method thus constructed can readily be applied to the independent component analysis. Its remarkable performance is illustrated by a numerical simulation.

Keywords

Cite

@article{arxiv.cs/0001004,
  title  = {Multiplicative Algorithm for Orthgonal Groups and Independent Component Analysis},
  author = {Toshinao Akuzawa},
  journal= {arXiv preprint arXiv:cs/0001004},
  year   = {2007}
}

Comments

11 pages, 2 figures