scglue.num.cov_mat

scglue.num.cov_mat(X, Y=None, bias=False)[source]

Covariance matrix (sparse friendly)

Parameters:

X (typing.Union[numpy.ndarray, scipy.sparse._matrix.spmatrix]) – First design matrix
Y (typing.Union[numpy.ndarray, scipy.sparse._matrix.spmatrix, None]) – Second design matrix (optional)
bias (bool) – Whether to return unbiased or biased covariance estimation

Return type:

numpy.ndarray

Returns:

cov – Covariance matrix, if only X is given. Cross-covariance matrix, if both X and Y are given.