pyriemann.datasets.make_gaussian_blobs

pyriemann.datasets.make_gaussian_blobs(n_matrices=100, n_dim=2, class_sep=1.0, class_disp=1.0, return_centers=False, center_dataset=False, random_state=None, centers=None, *, n_jobs=1, sampling_method='auto')

Generate SPD matrices for two classes.

Generate a set of SPD matrices drawn from Riemannian Gaussian distributions, one per class. Currently, it supports two classes. The distributions have the same dispersions. Useful for testing classification or clustering methods.

Parameters:

n_matricesint, default=100

Number of matrices to generate for each class.

n_dimint, default=2

Dimensionality of the generated SPD matrices.

class_sepfloat, default=1.0

Distance between the centers of the classes.

class_dispfloat, default=1.0

Dispersion of the matrices for each class.

centersNone | ndarray, shape (2, n_dim, n_dim), default=None

Centers for each class. If None, the centers are drawn randomly based on class_sep.

return_centersbool, default=False

If True, return the centers of each class.

center_datasetbool, default=False

If True, re-center dataset to the Identity. If False, dataset is centered around a random SPD matrix.

random_stateint, RandomState instance or None, default=None

Pass an int for reproducible output across multiple function calls.

n_jobsint, default=1

The number of jobs to use for the computation. This works by computing each of the class centroid in parallel. If -1 all CPUs are used.

sampling_method{“auto”, “slice”, “rejection”}, default=”auto”

Method used to sample eigenvalues: “auto”, “slice” or “rejection”. If “auto”, sampling_method will be equal to “slice” for n_dim != 2 and equal to “rejection” for n_dim = 2.

Added in version 0.4.

Returns:

Xndarray, shape (2*n_matrices, n_dim, n_dim)

Set of SPD matrices, for two classes.

yndarray, shape (2*n_matrices,)

Labels corresponding to each matrix.

centersndarray, shape (2, n_dim, n_dim)

The centers of each class. Only returned if return_centers=True.

Notes

Added in version 0.3.