ShICA-J
- shica.shicaj.shica_j(X_list, max_iter=10000, tol=1e-05, use_jointdiag=True, use_scaling=True, verbose=False)
- Parameters
X_list (ndarray of shape (m, k, n)) – input data
max_iter (int) – Maximum number of iterations to perform
tol (float) – Tolerance. The algorithm stops when the loss decrease is below this value.
verbose (bool) – If True, prints information about convergence
- Returns
W_list (ndarray of shape (m, k, k)) – Unmixing matrices
Sigmas (ndarray of shape (k,)) – Noise covariances
Y_avg (ndarray of shape (k, n)) – Source estimates
ShICA-ML
- shica.shicaml.shica_ml(X_list, max_iter=3000, init='shica_j', W_init=None, Sigmas_init=None, tol=1e-08, verbose=False)
- Parameters
X_list (ndarray of shape (m, k, n)) – input data
max_iter (int) – Maximum number of iterations to perform
init (None or "shica_j") – If “shica_j” uses shica_j to initialize unmixing matrices and noise covariance matrices. Parameters W_init and Sigmas_init are ignored. If None parameters W_init and Sigmas_init are used to initialize unmixing matrices and noise covariance matrices
W_init (ndarray of shape (m, k, k)) – Initial unmixing matrices
Sigmas_init (ndarray of shape (m, k)) – Initial noise covariances
tol (float) – Tolerance. The algorithm stops when the loss decrease is below this value.
verbose (bool) – If True, prints information about convergence
- Returns
W_list (ndarray of shape (m, k, k)) – Unmixing matrices
Sigmas (ndarray of shape (k,)) – Noise covariances
Y_avg (ndarray of shape (k, n)) – Source estimates