Fixed point ICA. Main algorithm of FASTICA.
[A, W] = fpica(whitesig, whiteningMatrix, dewhiteningMatrix, approach, numOfIC, g, finetune, a1, a2, mu, stabilization, epsilon, maxNumIterations, maxFinetune, initState, guess, sampleSize, displayMode, displayInterval, verbose);
the whitened data as row vectors
can be obtained with function whitenv
can be obtained with function whitenv
[ 'symm' | 'defl' ] the approach used (deflation or symmetric)
[ 0 - Dim of whitesig ] number of independent components estimated
[ 'pow3' | 'tanh' | 'gaus' | 'skew' ] the nonlinearity used
[same as g + 'off'] the nonlinearity used in finetuning.
parameter for tuning 'tanh'
parameter for tuning 'gaus'
step size in stabilized algorithm
[ 'on' | 'off' ] if mu < 1 then automatically on
stopping criterion
maximum number of iterations
maximum number of iteretions for finetuning
[ 'rand' | 'guess' ] initial guess or random initial state. See below
initial guess for A. Ignored if initState = 'rand'
[ 0 - 1 ] percentage of the samples used in one iteration
[ 'signals' | 'basis' | 'filters' | 'off' ] plot running estimate
number of iterations we take between plots
[ 'on' | 'off' ] report progress in text format
Perform independent component analysis using Hyvarinen's fixed point algorithm. Outputs an estimate of the mixing matrix A and its inverse W.