A New ICA Algorithm Based on Inner Product
Jiang Yu · 2005
Independent Component Analysis (ICA) is developed in recent years,which is a solution to Blind Source Separation problem basing on the high-order statistics features of signals.This paper presents a new approach basing on inner product of vectors to ICA,named I-ICA.I-ICA uses mixed data to cluster in new coordinates,eventually in- fers the mixed matrix A',and then computes the source signals through A'~(-1) and observed data.