A general class of ψ-APEX PCA neural algorithms

摘要

Principal component analysis (PCA) can be successfully applied to a variety of signal processing problems. Different analyzers have been reported in the scientific literature; among others, the Adaptive Principal component EXtractor (APEX) by Kung and Diamantaras has attracted much interest in the scientific community since it involves a specific neural architecture and a specific learning theory. The aim of this brief is to present a general class of APEX-like learning rules (referred to as ψ-APEX) and to illustrate their features by theoretical and numerical analysis.