the number of independent components is determined from the correlations between blocs issued from the dataset
res=ica_blocs_signals(x,n_ics,split,(options));
a matrix (n x q) or a Div structure
maximum number of independent components
choice of the method for splitting x, and n_b = the number of blocs
n_b (integer): random; n_b blocs
'jckn_b' (n_b=integer): Jack knife = contiguous blocs; n_b blocs; ex: 'jck5'
'vnbn_b' (n_b=integer): Venitian blinds; n_b blocs; ex: 'vnb5'
v (vector): a conjunctive vector of length n, identifying the observation to the blocs by the numbers 1 to n_b
an optional structure designed for the tuning fields of the ICA (see icascores help) and the field ibbplot for the figures:
options.ibbplot: 0 = no plot (by default); 1 = significant + lowest correlations; 2 = all
the obtained signals for i=1 to n_b blocs and j=1 to n_ics independent components
res.signal is a cell of dimensions (n_b x n_ics)
res.signal(i,j).entries is a Div structure
res.signal(i,j).entries.d is a vector or matrix of dimensions (q x j)
the scores obtained for i=1 to n_b blocs and j=1 to n_ics independent components
res.scores is a cell of dimensions (n_b x n_ics)
res.scores(i,j).entries is a Div structure
res.scores(i,j).entries.d is a vector or matrix with j columns and a number of observations depending on the splitting
the absolute value of the simple correlation coefficients between the signals obtained for the n_b blocs and j=1 to n_ics independent components
res.signal_corr is a cell of dimensions (1 x n_ics)
res.signal_corr(1,j).entries is a vector of dimensions ((a*j)^2,1)
the absolute value of the simple correlation coefficients between the scores obtained for the n_b blocs and j=1 to n_ics independent components
res.scores_corr is a cell of dimensions (1 x n_ics)
res.scores_corr(1,j).entries is a vector of dimensions ((n_b*j)^2,1)
number of ICs and minimal correlations; a Div structure
res.rmin.d is a matrix of dimensions (n_ics x 2)