This page is a reference documentation. It only explains the function signature, and not how to use it. Please refer to the user guide for the big picture.
Compute sparse precision matrices and covariance matrices.
The precision matrices returned by this function are sparse, and share a common sparsity pattern: all have zeros at the same location. This is achieved by simultaneous computation of all precision matrices at the same time.
Running time is linear on max_iter, and number of subjects (len(subjects)), but cubic on number of features (subjects.shape).
subjects : list of numpy.ndarray
alpha : float
max_iter : int, optional
tol : positive float or None, optional
verbose : int, optional
probe_function : callable or None
debug : bool, optional
emp_covs : numpy.ndarray, shape (n_features, n_features, n_subjects)
precisions : numpy.ndarray, shape (n_features, n_features, n_subjects)
The present algorithm is based on:
Jean Honorio and Dimitris Samaras. “Simultaneous and Group-Sparse Multi-Task Learning of Gaussian Graphical Models”. arXiv:1207.4255 (17 July 2012). http://arxiv.org/abs/1207.4255.