Rolf Unbehauen | Search PubMed

PubMed1999

Comments on: A unified algorithm for principal and minor components extraction.

This letter points out that the unified algorithm given by Chen et al. (Chen, T., Amari, S., & Lin, Q. (1998). Neural Networks, 11, 385-390) is a direct generalization of our invariant-norm algorithm. However, this direct-generalized unified algorithm is not practical from the learning point of view as the involved computations are intensive. As a matter of fact, a more effective generalization is made available.

Journal Article↗

PubMed1997

A Minor Component Analysis Algorithm.

The eigenvectors corresponding to the smallest eigenvalues of the autocorrelation matrix of the input signals are defined as the minor components, which play a very important role in many fields of adaptive signal processing such as spectral estimation, total least squares processing, eigen-based bearing estimation, digital beamforming, moving target indication, and clutter cancellation. This paper proposes a learning algorithm which extracts adaptively the minor component. We will use the Rayleigh quotient as an energy function and prove both analytically and by simulation results that the weight vector provided by the proposed algorithm is guaranteed to converge to the minor component of the input signals. Copyright 1997 Elsevier Science Ltd. All Rights Reserved.

Journal Article↗