ICA algorithm based on Hill Porter independence cr
no vote
Application background |--------------------| FAST; KERNEL; ICA; |--------------------|Version 1 - February 2007License see, below MPLPackage contains a Matlab implementation of the Fast Kernel ICA  This;As described in [1].  algorithm;ICA is based on minimizing a kernel measure of statistical KernelIndependence, the Hilbert-Schmidt norm of the covariance namelyIn feature space operator ([3] this: is called HSIC  ). See; an Given (nM matrix) W of n samples from m mixed sources the, goal is to find a xMatrix X such that the dependence between the estimated demixingSources X'*W is minimal.   FastKICA; uses an approximate Newton unmixedTo perfom this optimization.   For; more information on the methodAlgorithm, [1] and, for more information on HSIC refer, to [3]. readFunctions'chol_gauss'and'amariD' are taken from and based on The,Respectively, from Francis Bach code (at avai